An Efficient Web Service Composition Approach Based on improved Particle Swarm Optimization Model

Along with the rapid growth of the number of the available Web services over the Internet, the quality of services (QoS) becomes an imperative concern for service consumers and providers. How to differentiate their quality based on Qos criteria at Web service level is a challenging issue. Considering that evolutionary algorithm has powerful ability of searching information, a new Web service composition model under improved particle swarm optimization (IPSO) algorithm is presented in this paper. According to the non-functional attributes Qos of services, the model can select Web services with optimal Qos and compose them into value-added Web services. Furthermore, IPSO model not only can find the service with optimal QoS immediately but also can overcome the bottleneck of the standard PSO model. It can balance the exploitation and exploration and avoid premature convergence to some extent. As a result, web services can be invocated dynamically and be composed automatically and efficiently. The preliminary experimental results indicate that the Web service composition approach based on IPSO model is applicable and feasible to the optimization of general service compositions.