Research And Application Of Dynamic Web Service Composition Based On Ontology

Web service is a kind of distributed computing model, which develops very prosperously recently; it can realize the seamless operation between the application and application groupware in different platform and different programming language. With the perfect of the web service criterion and the development of its platform, more and more services are becoming available on the internet, and also the applications based on the Web service; but most of them are single and monotony, which can not meet the actual application demand very well. We need to compose various Web services to complete a complex task. Traditional method is to choose some suitable Web service artificially based on the demand, then define the interactive process between these services, and complete the composition of Web service through hard code as the last step. But with the increasing of Web service, it is unpractical to compose the services manually, and the quality of service can not be ensured either. So we must compose web services dynamically.Dynamic Web services composition means that the program or agent can pick and assemble suitable Web service to complete the task automatically, while just need a standard description given by the customer from higher level.This thesis is focus on the dynamic service composition based on the semantic Web, the aim is to compose the various Web services automatically according to the demand of customers, and to complete the task corporately; this can make the services meet the demand of the customers neatly.In this thesis, first of all, a kind of service matching strategy based on reasoning and semantic similarity computing had been bought out, and a service composition algorithm of backward matching was bought out as following, and gave the method of execution and exception handle. At the end of the thesis, a model of dynamic service composition had been designed and realized partly, and domain ontology and service ontology for experiment had been built. From the experimental results, we can discover that this model can compose web service very efficiently and correctly based on the output and input.