ABSTRACT
Recent work in genetic programming (GP) has highlighted the need for stronger benchmark problems. For benchmarking planning scenarios, the artificial ant problem is often used. With a limited number of test cases, this problem is often fairly simple to solve. A more complex planning problem is Tartarus, but as of yet no standard representation for Tartarus exists for GP. This paper examines an existing parse tree representation for Tartarus, and identifies weaknesses in the way in which it manipulates environmental information. Through this analysis, an alternative representation is proposed for Tartarus that shares many similarities with those already used in GP for planning problems. Empirical analysis suggests that the proposed representation has qualities that make it a suitable benchmark problem.
- D. Ashlock and J. Freeman. A pure finite state baseline for tartarus. In Proceedings of the 2000 Congress on Evolutionary Computation, pages 1223--1230, 2000.Google ScholarCross Ref
- D. Ashlock and M. Joenks. ISAc lists, A different representation for program induction. In J. R. Koza, W. Banzhaf, K. Chellapilla, K. Deb, M. Dorigo, D. B. Fogel, M. H. Garzon, D. E. Goldberg, H. Iba, and R. Riolo, editors, Genetic Programming 1998: Proceedings of the Third Annual Conference, pages 3--10, University of Wisconsin, Madison, Wisconsin, USA, 1998. Morgan Kaufmann.Google Scholar
- D. Ashlock and M. McRoberts. A gp-automata reprise of astro teller's bulldozer experiment. Technical report AM97--17, ISU Mathematics, 1998.Google Scholar
- D. Ashlock and E. Warner. The geometry of tartarus fitness cases. In J. Wang, editor, 2008 IEEE World Congress on Computational Intelligence, Hong Kong, 2008. IEEE Computational Intelligence Society, IEEE Press.Google ScholarCross Ref
- D. Ashlock, S. Willson, and N. Leahy. Coevolution and tartarus. In Proceedings of the 2004 IEEE Congress on Evolutionary Computation, pages 1618--1624, Portland, Oregon, 2004. IEEE Press.Google ScholarCross Ref
- M. C. J. Bot, N. Urquhart, and K. Chisholm. Agent motion planning with gas enhanced by memory models. In L. Spector, E. D. Goodman, A. Wu, W. Langdon, H.-M. Voigt, M. Gen, S. Sen, M. Dorigo, S. Pezeshk, M. H. Garzon, and E. Burke, editors, Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2001), pages 227--234, San Francisco, California, USA, 2001. Morgan Kaufmann.Google Scholar
- G. Cuccu and F. Gomez. When novelty is not enough. In C. Chio, S. Cagnoni, C. Cotta, M. Ebner, A. Ekárt, A. I. Esparcia-Alcázar, J. J. Merelo, F. Neri, M. Preuss, H. Richter, J. Togelius, and G. N. Yannakakis, editors, Applications of Evolutionary Computation, volume 6624 of Lecture Notes in Computer Science, pages 234--243. Springer Berlin Heidelberg, 2011. Google ScholarDigital Library
- F. J. Gomez. Sustaining diversity using behavioral information distance. In Proceedings of the 11th Annual conference on Genetic and evolutionary computation, GECCO'09, pages 113--120, New York, NY, USA, 2009. ACM. Google ScholarDigital Library
- D. Kim. Analyzing sensor states and internal states in the tartarus problem with tree state machines. In X. Yao, E. Burke, J. A. Lozano, J. Smith, J. J. Merelo-Guervós, J. A. Bullinaria, J. Rowe, P. T. A. Kabán, and H.-P. Schwefel, editors, Parallel Problem Solving from Nature - PPSN VIII, volume 3242 of Lecture Notes in Computer Science, pages 551--560, Birmingham, UK, 2004. Springer-Verlag.Google Scholar
- J. R. Koza. Genetic Programming: On the Programming of Computers by Means of Natural Selection. MIT Press, Cambridge, MA, USA, 1992. Google ScholarDigital Library
- J. R. Koza. Genetic Programming 2 - Automatic Discovery of Reusable Programs. MIT Press, Cambridge, MA, USA, 1994. Google ScholarDigital Library
- J. McDermott, U.-M. O'Reilly, S. Luke, and D. White. gpbenchmarks.org -- problem classification (control/planning problems). http://www.gpbenchmarks.org/wiki/index.php?title=Problem_Classification%#Control.2FPlanning_problems, Sept. 2012.Google Scholar
- J. McDermott, D. R. White, S. Luke, L. Manzoni, M. Castelli, L. Vanneschi, W. Jaskowski, K. Krawiec, R. Harper, K. De Jong, and U.-M. O'Reilly. Genetic programming needs better benchmarks. In Proceedings of the fourteenth international conference on genetic and evolutionary computation, GECCO'12, pages 791--798, New York, NY, USA, 2012. ACM. Google ScholarDigital Library
- D. J. Montana. Strongly typed genetic programming. Evol. Comput., 3(2):199--230, June 1995. Google ScholarDigital Library
- A. Teller. The evolution of mental models. In K. E. Kinnear, Jr., editor, Advances in Genetic Programming, chapter 9, pages 199--219. MIT Press, 1994. Google ScholarDigital Library
- A. Trenaman. Concurrent genetic programming, tartarus and dancing agents. In R. Poli, P. Nordin, W. B. Langdon, and T. C. Fogarty, editors, Genetic Programming, Proceedings of EuroGP'99, volume 1598 of Lecture Notes in Computer Science, pages 270--282, Goteborg, Sweden, 1999. Springer-Verlag. Google ScholarDigital Library
- P. A. Whigham. Grammatically-based genetic programming. In J. P. Rosca, editor, Proceedings of the Workshop on Genetic Programming: From Theory to Real-World Applications, pages 33--41, Tahoe City, California, USA, 9 July 1995.Google Scholar
- P. Zahadat and S. D. Katebi. Tartarus and fractal gene regulatory networks with inputs. Advances in Complex Systems, 11(06):803--829, 2008.Google ScholarCross Ref
Index Terms
- An effective parse tree representation for tartarus
Recommendations
A true finite-state baseline for tartarus
GECCO '13: Proceedings of the 15th annual conference on Genetic and evolutionary computationThe Tartarus problem is a benchmark problem for non-Markovian decision making. In order to achieve high fitness, individuals must make efficient use of internal state. Finite-state machines are an ideal candidate for exploring the Tartarus problem, and ...
Adaptable representation in GP
GECCO '05: Proceedings of the 7th annual workshop on Genetic and evolutionary computationGenetic Programming uses trees to represent chromosomes. The user defines the representation space by defining the set of functions and terminals to label the nodes in the trees. The sufficiency principle requires that the set be sufficient to label the ...
Evaluation of crossover operator performance in genetic algorithms with binary representation
ICIC'11: Proceedings of the 7th international conference on Intelligent Computing: bio-inspired computing and applicationsGenetic algorithms (GAs) generate solutions to optimization problems using techniques inspired by natural evolution, like crossover, selection and mutation. In that process, crossover operator plays an important role as an analogue to reproduction in ...
Comments