Research On Web Service Discovery Framework Based On OWL-S

In recent years,with the development of e-commerce,the requirement of translationable is more and more integratable.Web services as the mainstream techmology of distributed computing and Web application,provide an effective solution for the problems of heterogeneous systems interoperability,mutual communication.However, the existing UDDI Web service discovery framework is based on keyword arch,which means that traditional UDDI query does not have a semantic search capabilities. To solve this problem, if we introduce the semantic Web technology into Web services,when using this service,we could use semantic reasoning semantic matching of Web services. Web services Ontology Language OWL-S is the bridge between the Semantic Web and the Web services.OWL-S could descripe and reason Web services,which provid conditions for the semantic matching.Based on the analysis of the traditional Web services and Semantic Web,this paper constructed the UDDI-based Semantic Web services model based on the hierarchical matching model. The first layer is services matched layer,this layer matches the services through the application field,eliminates the services not belong to the same application service.The second layer is PE matching layer, through this layer, you can filter out the implementation which does not meet the conditions of service and implementation effectivenes of services.the final layer is IO matched layer, which is filter those services with a semantic similarity below the specified value of the service. Through screening of the first and the second layers, reduced the IO match the workload. The third layer is the core of the matching model. The method of establishment is: by through analysis of the concept of similarity algorithms, a matching algorithm which is based on geometric distance and concept of information combination similarity is to be found; to find the best match between input and output in requested service and advertising services, the matching problem has been converted into the IO assignment problem, and proposed the Hungarian assignment algorithm.Finally a standard test is conducted by using MyEclipse6.0, JDK6.0,some source-opening resources softwares and the OWLS-TC2 which is developed by the German Research Center for Artificial Intelligence, the experiments illustrate that the Matching algorithm is feasible and can achieve a more precise matching service.