Research On Verification Of Web Services Composition And Services Selection Algorithm

Web services make the enterprise IT architecture more powerful in interoperability, flexibility and reusability. In addition, the cooperation methods between commercial partners are changed by using Web services in some sense. Through services composition, companies could share resources and automate business flow. So Web services composition research is becoming a hot area during these years.Conjunction of services choreography and services orchestration could enhance efficiency and veracity of the business process modeling greatly. WS-BPEL and WS-CDL are the main services composition specifications that based on services choreography and services orchestration respectively. However, since they belong to different technical specification levels, it is better to maintain the consistency of their behaviors. How to verify consistency of services choreographs and orchestration is a problem need to be solved.As more and more Web services are developed, the issue large-scale Web services selection has more weight than before. The key point of this issue is how to propose an efficient algorithm for this issue. However, most work on hand has not addressed this.To solve the problem above, points of the thesis are mainly on the following:1) CSP-based verification for services choreography and services orchestration. According to the translation rules proposed, both WS-CDL and WS-BPEL are mapping into CSP language, then model checking tool PAT, which is based on CSP, check the consistency. We will give an e-business case in our paper.2) An efficient algorithm GAELS proposed for large-scale Web services selection. Aiming to service selection problem with mass candidates, we propose a new algorithm named GAELS (Genetic Algorithm Embedded Local Searching), which uses the strategies of high-fitness initial population and mutation with local searching to speed the convergence. The in-depth experimental results show that the proposed algorithm GAELS can get the approximately optimal solution more quickly and be more adaptive to the expanding of candidate services than simple genetic algorithm.