A Novel Approach to Large-Scale Services Composition
We investigate a multi-agent reinforcement learning model for the optimization of Web service composition in this paper. Based on the model, a multi-agent Q-learning algorithm was proposed, where agents in a team would benefit from one another. In contrast to single-agent reinforcement-learning, our algorithm can speed up the convergence to optimal policy. In addition, it allows composite service to dynamically adjust itself to fit a varying environment, where the properties of the component services continue changing. A set of experiments is given to prove the efficiency of the analysis. The advantages and the limitations of the proposed approach are also discussed.
KeywordsWeb Service composition multi-agent
Unable to display preview. Download preview PDF.
- 2.Carman, M., Serafini, L., Traverso, P.: Web service composition as planning. In: ICAPS 2003 Workshop on Planning for Web Services, pp. 1636–1642 (2003)Google Scholar
- 3.Doshi, P., Goodwin, R., Akkiraju, R., Verma, K.: Dynamic workflow composition using markov decision processes. In: IEEE International Conference on Web Services, pp. 576–582. IEEE (2004)Google Scholar
- 5.Gonzaga, T., Bentes, C., Farias, R., de Castro, M., Garcia, A.: Using distributed-shared memory mechanisms for agents communication in a distributed system. In: Seventh International Conference on Intelligent Systems Design and Applications, ISDA 2007, pp. 39–46. IEEE (2007)Google Scholar
- 7.Kaelbling, L., Littman, M., Moore, A.: Reinforcement learning: A survey. Arxiv preprint cs/9605103 (1996)Google Scholar