Research On QoS-based Trusted Web Service Composition Optimization Method

Web service composition technology can effectively integrate distributed single-function Web services on the Internet and achieve richer business logic and functions by constructing composite services.The composition process could be divided into two key steps:combination design and service optimization.In fact,while the functional attributes are satisfied,users will pay more attention to the user experience,that is,quality of service.However,in a complex network environment,the quality of service of Web service is usually in dynamic change.The Web service and the credibility of QoS published by service providers need to be verified.Traditional QoS-based service composition method merely constructs the service portfolio from the perspective of optimal service quality and cannot guarantee that the service portfolio is true and reliable.Firstly,to solve the problem of Web service composition optimization in complex network environment,the shortcomings of the existing combination strategy and optimization algorithm are analyzed and based on the combination of credibility evaluation and QoS-based combination strategy,a QoS-based trusted Web service composition strategy is proposed in this paper.Then,based on the historical behavior of Web service,the credibility of Web service published by service provider is evaluated from both the objective and subjective aspects.In objective aspect,the Bayesian learning theory is used for estimation based on the service call record,and in subjective aspect,the historical user feedback information is used for evaluation.Based on the credibility evaluation of Web services and the structure of service composition workflow,a calculation method for credibility evaluation of Web service composition is proposed in this paper.The results of Web service credibility evaluation simulation experiment shows that the credibility evaluation model proposed in this paper could reasonably and comprehensively evaluate the credibility of Web service and avoid the one-sidedness brought by historical user evaluation method.Finally,using the trusted Web services QoS attributes after credibility measurement,and considering the conflict among QoS attributes,a multi-objective optimization model of Web service composition with constraint information is constructed in this paper.In order to improve the optimization efficiency and optimization effect of QoS in Web service composition,an improved multi-objective grey wolf optimizer is proposed.The algorithm uses the good point set theory to initialize the population and ensure the rich diversity of the population in the initial stage of algorithm optimization,and uses nonlinear control parameter adjustment strategy coordination algorithm exploration and development ability to achieve better optimization performance.In order to verify the effectiveness of the proposed algorithm,the QoS optimization process of Web service composition on QWS dataset and Simulation dataset is compared in this paper,and the results shows that the improved multi-objective grey wolf optimizer could solve the problem of Web service composition optimization well,and it owns faster optimization speed and better optimization results than similar methods.