ISPCA: IDPSO-Based Service Pool Construction Algorithm

Abstract

QoS-driven web services selection plays an important role in web service composition. Since the selected services would fail down because of the dynamic network situation, this paper presents a service pool model to help users save more composite plans in one selection process. In addition, an efficient service pool construction algorithm, ISPCA, based on an Improved Discrete Particle Swarm Optimization (IDPSO) is also proposed. And the third contribution of our work was that a non-uniform mutation strategy (NUMS) is introduced to the global best particle in IDPSO to enhance the quality of composite plans in service pool. The results of experimental evaluation indicate that our approach significantly outperforms existing methods in execution time while selecting more combinations with better QoS performance.