基于梯度的改进动态Web服务选择算法

An Improved Dynamic Web Services Selection Algorithm Based on Gradient

随着Web2.0的迅速发展,互联网上发布的Web服务越来越多,不同服务供应商提供的服务通过整合以提供功能更强大的组合服务。每个服务节点上功能相似的Web服务的QoS(quality of service)不同,因此,QoS全局最优动态Web服务选择成为了服务组合中的一大挑战。在传统的粒子群优化算法的基础上引入梯度的思想,文中设计了一种用于解决动态Web服务选择问题的改进算法gPSO-GODSS。将问题抽象为带QoS约束的多目标组合优化问题,并进一步将其向单目标转化。利用梯度的方法改进粒子群算法的更新速度,从而改进算法的收敛速度,最终产生一组满足约束条件的优化服务组合流程集。理论分析和实验结果证明了该算法的可行性和有效性,且gPSO-GODSS算法收敛的执行效率和收敛速度均优于已有的PSO-GODSS算法。

As the rapid development of Web2.0,there are more and more Web services on the Internet.Web services from different service providers can be integrated to form a composited service.As the Web services on each node with similar functions have different QoS (quality of service),dynamic Web service selection with global QoS optimization becomes a critical issue in Web service composition.In order to solve the problem,based on the basic particle swarm optimization (PSO) and the thought of gradient,we propose a gPSO-GODSS (global optimal of dynamic Web services selection based on PSO with gradient).We abstract the original Web service selection problem into a multi-objective services composition optimization with global QoS constraints,which is further transformed into a single-object.The gradient method is used to improve the update speed of basic PSO,thus improving the convergence speed of the PSO-GODSS algorithm.Then,the improved PSO is exploited to produce a set of optimal services composition process with QoS constraints.Theoretical analysis and experimental results indicate the feasibility and efficiency of this algorithm,and the execution efficiency and convergence rate of gPSO-GODSS algorithm are both better than the existing PSO-GODSS algorithm.