Research And Application Of Web Service Selection Based On QoS

This paper focuses on the selection of Web services,and studies the comprehensive evaluation and service selection algorithm of Web services from the perspective of quality of service(QoS).Web service is an implementation of SOA architecture.The emergence of SOA architecture enables the system to be integrated and invoked in the form of services,thus realizing the rapid development of Internet Web Application System.However,with the increasing number of Web services published on the network,there are many services with the same functions but different non-functional attributes,and some services need to be combined because it only provide a single function.How to evaluate these services comprehensively and how to select the service that can meets the needs in service composition has become a hot issue in service selection.For the comprehensive evaluation of Web services based on QoS,the existing methods mainly include subjective weight,objective weight and subjective and objective weight.In this paper,it is adopts a dynamic comprehensive evaluation method based on subjective and objective weights.The G1(Sequential Relation Analysis)method is used to calculate subjective weights.Developers only need to provide the order relationship and importance between the attributes of QoS,and the system can calculate the subjective weights.The objective weights are calculated by the variation coefficient method.The dynamic changes of Web services from the perspective of time series are analyzed,and Web services are evaluated comprehensively by reward(punishment)mechanism.Aiming at the choice of Web services,this paper proposes to transform four basic workflow structures into hybrid workflow based on sequential and parallel structure,and gives the calculation method of each attribute under this workflow.At the same time,service selection problem based on hybrid workflow is modeled and solved by improved genetic algorithm.In the improved genetic algorithm,the idea of niche and concentration factor is used to improve the speed and convergence of the algorithm.Finally,an experimental system of Web service selection based on QoS is designed and implemented in this paper.The system can calculate the quality of service attributes and the objective weight of each attribute for a period of time according to the information of the service being called.The subjective weight can be calculated according to the requirement of developers for service quality.Finally,the comprehensive evaluation and ranking of Web services are given.At the same time,the system supports the process of service composition.According to the composite workflow diagram provided by developers,Web services satisfying the quality of service requirements can be selected from the Web service system.