QoS-based Web Service Composition

Web service has been rapidly developed as a new Web pattern in recent years. How to integrate varaious Web services into a newly value-added and complex service to meet the requirenment of different users is a popular research area. At present most of the research on QoS-based Web service composition is local optimum. These methods are usually only meet the functional requirements of earch service node in the service composition process, can not solve the service composition with QoS global optimization problems. The main research work can be attributed to the following aspects:Firstly, establish the QoS-based Web service compositon framework. After studying on the existing Web services architecture, combined the characteristics of service composition with requirements of QoS, a QoS-based web service composition framework was proposed.users can Customized services in the service composition process, and can give a service evaluation feedback to the system. Besides, a QoS adaptation is added in the framework. The framework not noly can support QoS better and reflect the user’s preferences, but also have good scalability.Secondly, design a QoS aggregation algorithm based on m-express.in Service composition, Web service composition model is formalized into an m-express. The m-express is conducive to the QoS indicators aggregation in Web service composition. What’s more, in the case of comprehensive consideration of the user’requirements of QoS, for five common QoS evaluation indicators, an m-expression-based QoS aggregation algorithm is presented. The aggregation algorithm is better to assess the quality of service composition, and better to meet the needs of users.Thirdly, Web services selection with QoS global optimal in Web services composition is presented.Particle swarm algorithm to solve the service composition problem have a great advantage, but the particle swarm algorithm prone to premature convergence into local optimum. By introduceing chaos search and local search into particle swarm algorithm, proposed a Web services selection with QoS global optimal in Web services composition. Particle Swarm update process is divided into two phases, the previous stage into chaos search to prevent the local convergence of the algorithm; later stage to introduce local search strategy, the solution tends to the forefront of the Pareto solution.’Finaly, design and implement a QoS-based Web service composition prototype system. A prototype system based on the integration of previous studies, including the release of the service, registration, query and combination modules. Service selection algorithm and verify the feasibility and effectiveness.