Research On Semantic Web Service Discovery

With the rapid development of web, WWW has been a huge global informatiom repository. The Web Service paradigm is transforming the Web from a provider of static pages to a provider of interactive, automated and intelligent services that interact via the Internet. In order to employ its full potential, the web service paradigm must be supported by an appropriate discovery mechanism. Current Web Services discovery technology is based on UDDI and WSDL. However, UDDI technology is based on the keyword-based discovery method. The shortcomings of keyword-based techniques are their low precision and recall. Many irrelevant services may include in their description the query keywords, leading to low precision. Also, the query keywords may be semantically equivalent but syntactically different from the words in the offered services, leading to reduced recall. The key underlying problem is that keywords are a poor way to capture the semantics of a service request or service advertisement.The targets of Web Services discovery are high efficiency and automatization. So, the services discovery technologys in Web Service (such as UDDI) cannot satisfy the requirement. Using semantic description and service ontology are available approaches on service discovery. How to describe the Web service and how to make the semantic Web, ontology and semantic inference apply to the finding and matching in Web Service are the problems which need to be further researched.In this thesis,we first introduce the knowledge background of Web Service, service discovery mechanism and semantic web etc.Then we combine the semantic web and Web Service discovery and propose a matching algorithm. This algorithm achieves the rapid matching of requestor and provider adopting the idea of multilayer filter. At the same time, we lay a strong emphasis on how to realize the combination of several services to match the request. Finally, we explain the implementation of the algorithm by using instances.