QoS-constrained Evaluation And Selection Of Web Service In Information Support System

Service-oriented architecture(SOA)and Web service technology are widely used in the military field.Interaction between services based on Web in command and control system has been a trend.Web service realize SOA,which use standard protocols to ensure interaction between services on different platforms.The combination of Web services has become the focus of academia and industry because of its ability to realize the reuse and appreciation of services.The command and control system of Service-oriented architecture should provide the service that has reasonable function and good performance for decision makers.And the Quality of Service(QoS)directly determines the availability and effectiveness of military information services.When regarding QoS constraints as essential,the process has a diffucultiy selecting combined service from a big candidate space.So optimizing process is primary.In addition,the existing service selection methods are assumed the QoS value is constant,but in practice,the performance of Web services is variational with the change of service environment.So the effective prediction method is necessary.This thesis mainly focuses on several key technical issues of QoS-constrained web service select.The main tasks include:(1)The Skyline method in database query is introduced.And the representative Skyline service selection method based on the dominance relationship is proposed.The main idea is making a classification in the whole set of services according to the various QoS attribute dimensions,and choosing typical Web services to build a binary tree,and then applying recursion traversal the tree to find a service that meets global QoS constraints of the optimal service composition.(2)The method of QoS prediction based on time series is proposed.The uncertain analysis of the QoS is a precondition,and history QoS data is used to establish the QoSstate transition rules,and then the QoS value is determined depends on the time series model.This method has good feasibility and high prediction accuracy.(3)The QoS-constrained information Web service selection system is designed and developed.The web service model based on QoS is improved on the SOA model,which can complete the interaction between services and achieve the purpose of service selection.