| There is a trend of smart devices or enable macheines, which use of Internet of Things(IoT) as a platform to coordinate and communicate for realizing intelligent decision and control. Nowadays, IoT creares a new oppprtunity to build smart devices that better integrate of industry. On the basis of IT technologies, IoT contributes to develop management, improve production efficiency and reduce costs. With the proliferation of Web services providing similar functionality, how to construct synthesizing web services according to the non-functional properties, such as Quality of Service(QoS) has becomes a prominent issue. Therefore, in view of the real time and dynamic of IoT, this thesis contributes to the QoS-oriented web service compositon.This study presents a multi-criteria programming(MCP) model is developed to solve the QoS-oriented web service composition problem. Besides, a multi-criteria goal programming(MCGP) model is proposed to discover feasible composite service by relaxing QoS constraints if no suitable or available web service could fulfill user requirements. Moreover, genetic algorithm(GA) is developed to solve the presented MCGP model. To improve the convergence, local search ablility and prematurity of GA, multi-population genetic algorithm(MGA) is a form of multi-population with parallel evolution that gives equal importance to low fit individuals. This allows MGA to maintain the diversity of individuals in the population, and then ensures faster convergence.To compare MGA with GA, experimental results validate the excellent performance of MGA in terms of excellent convergence ability and powerful searching ability. It has economic value for large scale web service composition optimization problem. |
PDF Full Text Request