A GP Approach to QoS-Aware Web Service Composition and Selection

Created by W.Langdon from gp-bibliography.bib Revision:1.7615

@InProceedings{daSilva:2014:SEAL,

• author = "Alexandre {Sawczuk da Silva} and Hui Ma and Mengjie Zhang",
• title = "A {GP} Approach to {QoS}-Aware Web Service Composition and Selection",
• booktitle = "Proceedings 10th International Conference on Simulated Evolution and Learning, SEAL 2014",
• year = "2014",
• editor = "Grant Dick and Will N. Browne and Peter Whigham and Mengjie Zhang and Lam Thu Bui and Hisao Ishibuchi and Yaochu Jin and Xiaodong Li and Yuhui Shi and Pramod Singh and Kay Chen Tan and Ke Tang",
• volume = "8886",
• series = "Lecture Notes in Computer Science",
• pages = "180--191",
• address = "Dunedin, New Zealand",
• month = dec # " 15-18",
• publisher = "Springer",
• keywords = "genetic algorithms, genetic programming",
• isbn13 = "978-3-319-13562-5",
• DOI = "doi:10.1007/978-3-319-13563-2_16",
• abstract = "Web services are independent functionality modules that can be used as building blocks for applications that accomplish more specific tasks. The large and ever-growing number of Web services means that performing this type of Web service composition manually is infeasible, which leads to the exploration of automated techniques to achieve this objective. Evolutionary Computation (EC) approaches, in particular, are a popular choice because they are capable of efficiently handling the complex search space involved in this problem. Therefore, we propose the use of a Genetic Programming (GP) technique for Web service composition, building upon previous work that combines the identification of functionally correct solutions with the consideration of the Quality of Service (QoS) properties for each atomic service. The proposed GP technique is compared with two PSO composition techniques using the same QoS-aware objective function, and results show that the solution fitness values and execution times of the GP approach are inferior to those of both PSO approaches, failing to converge for larger datasets. This is because the fitness function employed by the GP technique does not have complete smoothness, thus leading to unreliable behaviour during the evolution process. Multi-objective GP and the use of functional correctness constraints should be considered as alternatives to overcome this in the future.",

}

Genetic Programming entries for Alexandre Sawczuk da Silva Hui Ma Mengjie Zhang

Citations