Automated Web service composition, which refers to the creation of a complex application from pre-existing building blocks (Web services), has been an active research topic in the past years. The advantage of having an automated composition system is that it allows users to create new applications simply by providing the required parameters, instead of having to manually assemble the services. Existing approaches to automated composition rely on planning techniques or evolutionary computing (EC) to modify and optimise composition solutions directly in their tree/graph form, a complex process that requires several constraints to be considered before each alteration. To improve the search efficiency and simplify the checking of constraints, this work proposes an indirect Particle Swarm Optimisation (PSO)-based approach. The key idea of the indirect approach is to optimise a service queue which is then decoded into a composition solution by using a planning algorithm. This approach is compared to a previously proposed graph-based direct representation method, and experiment results show that the indirect representation can lead to a greater (or equivalent) quality while requiring a lower execution time. The analysis conducted shows that this is due to the design of the algorithms used for building and evaluating the fitness of solutions.
CITATION STYLE
da Silva, A. S., Mei, Y., Ma, H., & Zhang, M. (2016). Particle swarm optimisation with sequence-like indirect representation for web service composition. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9595, pp. 202–218). Springer Verlag. https://doi.org/10.1007/978-3-319-30698-8_14
Mendeley helps you to discover research relevant for your work.