Font Size: a A A

Research On Web Service Composition Based On Improved Discrete Particle Swarm Optimization Algorithm

Posted on:2017-03-06Degree:MasterType:Thesis
Country:ChinaCandidate:Z H JingFull Text:PDF
GTID:2308330485964001Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
With the constantly-developing Internet technology, web service obtains the rapid development and wide applications as an emerging application schema in recent years. The number of web services grows increasingly. Meanwhile, as the growing users’demands for service function and quality of service (QoS), a single service has been unable to meet the users’increasingly complex business needs. Web service composition technology arises at the historic moment. Web service composition reuses the existing services distributed in Internet and realizes the seamless integration between services, thereby constructing the value-added applications to meet users’ complex demands. However, in view of substantial web services with identical functionality but different QoS on the network, how to select the services satisfying users’ needs from large-scale candidate services set efficiently and forms the value-added services, has become new application requirements and research hotspot.The paper solves the optimization problem of web service composition based on the modified discrete Particle Swarm Optimization Algorithm. It not only researches the traditional web service composition optimization problem in serial mode but also studies large-scale web service composition optimization problem in MapReduce parallel mode.The main research work is as follow:(1) Introduces the basic knowledge and related technology of web service composition, including the basic concept of web service and web service composition, the existing typical research methods in the service composition field and the relative basic theory of service composition based on QoS.(2) Analyses the basic principle of the standard PSO and points out the defect that the algorithm has premature convergence and is easy to trap into local optimum at later stages. This paper proposes an improved chaotic particle swarm optimization algorithm to solve service composition optimization problem. It introduces the idea of chaos optimization to improve the swarm diversity while simultaneously reduces the search space and improves the computational efficiency by skyline technology. A large number of experiments verify the effectiveness and stability of ICPSO for solving the web service composition optimization problem.(3) Based on MapReduce framework and basic principle, this paper models PSO by utilizing distributed parallel framework and presents a distributed parallel Particle Swarm Optimization Algorithm based on MapReduce (MR-IDPSO) combined with premature convergence processing mechanism, in order to solve large-scale service composition optimization problem in distributed parallel environment. The experimental results manifest that MR-IDPSO is feasible and effective for solving large-scale web service composition optimization problem in MapReduce mode.
Keywords/Search Tags:service composition, particle swarm optimization (PSO), quality of service (QoS), MapReduce, Skyline
PDF Full Text Request
Related items