Research On Web Service Composition Mechanism Based On Quantum Particle Swarm Optimization

With the rapid development of Services Computing technology and the increasing popularity of complex network application environments such as cloud computing, a traditionally single Web service cannot meet the users’ requirements any more. Complex Web applications usually need the combination of multiple Web services to work collaboratively. However, the increase of types and the number of Web services makes selection and composition of Web services more complicated than ever. Besides, there are lots of similar candidate Web services on the Internet and each of them can meet the users’ basic requirements, which makes it more and more difficult for service requesters to select and compose services. Therefore, how to select appropriate Web services and realize efficient and effective service composition based on their non-functional properties has become one of the hot topics.This thesis makes a comprehensive investigation on service composition and swarm intelligence optimization algorithms. It endeavors to increase the efficiency of complex service composition by employing swarm intelligence optimization methods. Firstly, a critical analysis on traditional particle swarm algorithms is made in this thesis. According to the analysis, an Improved Quantum Particle Swarm Optimization algorithm(IQPSO) is secondly proposed and implemented to address the problem of slow convergence and limited local search ability. Simulation results demonstrate the effectiveness and efficiency of IQPSO. Based on the proposed IQPSO, this thesis thirdly applies swarm intelligence optimization methods into service composition and presents an IQPSO-based Web Service Composition(IWSC) scheme. Simulation results show that compared with other representative schemes, IWSC has better performance in stability, convergence speed and success rate. Finally, this thesis gives a designing and implementation of a distributed Web Services composition simulation system based on IWSC. The simulation system is built on Model View Controller(MVC) and the praticality of IWSC is verified through the establishment of the system structure and speficication of all the included models. In comparison with the centralized Web service composition simulation system, the presented distributed system in this thesis has more advantages in meeting the demand of users and decreasing the requiring time.