Research On Composite Service Selection Method Supporting Semantic Matching And QoS

With the widespread application and development of service computing technology, the numberof service continues to grow explosively. There are numerous services with the same function butdifferent non functional properties in the internet, how to choose a group of services which cansatisfy the actual needs of the user from these services has become a hot issue in web serviceresearch. Most of the existing service selection methods don’t achieve the desired results. On theone hand the quality of service is uncertain due to the dynamic environment, services exhibitdifferent quality under different execution environment. On the other hand there is composabilityproblem between services. In this paper, its contents include the following areas:Firstly, we used cloud model theory to measure the uncertainty of the quality of candidateservices. We evaluate the comprehensive uncertainty of service based on the weights which is givento each quality attribute by users. We present a service filtering method based on the uncertainty ofQoS in order to provide optimized candidate services resources to service selection process. Thismethod can ensure that the services we provide to the user in the service selection process is stableand not affected by dynamic service environment.Secondly, to solve the problems of low interoperability and composability of services, wepropose a method to compute semantic matching degree of service composition base on bipartitegraph. This method is used for semantic analysis of combination services to ensure the selectedservices connected seamlessly and executed correctly.Thirdly, we propose a composite service selection method supporting semantic matching andthe quality of service, we evaluate the overall quality of combination service combined with thesemantic matching degree and non-functional attributes (QoS). Services selection is transformed tomulti-constrained and multi-objective optimization problem. GA is used to solve the problem ofcomposite service selection.Finally, we designed a composite service selection simulation system which supportingsemantic matching and QoS. Simulation results show that the composite service we provide to userby the method in this paper can meet the needs of users and execute correctly.