Research On Semantic-based Web Services Discovery

As the rapid development of Internet, there are more and more web services available on the web. However, how to discover useful web service in an efficient way becomes an important problem. In the traditional way, discovering a web service mainly based on service matching of UDDI architecture, but UDDI registration center can only provide query service through key-words and can not support semantic information. As a result, service matching is rather a kind of key-words and Syntax matching than a service matching; it makes for low recall ratio and precision ratio which can not satisfy the users. Further more, WSDL is a kind of service description language on the semantic level, it provides some physical information of technical details when invoking a certain web service, so it can not provide functional information because there is no functional descriptions.Semantic Web hammers at the semantization of web information, it can describe the information in a formal way, and then the information can be processed automatically by computer. The use of semantic web makes it possible to implement a matching method based on service function. This paper uses OWL-S as the description language which is ontology based that can describe web services on semantic web, in that way, the disadvantage of using WSDL for semantic can be remedied.The paper makes a deep research on web services and service discovering mechanism, semantic web, ontology, OWL and other related techniques, discusses the usage of semantic web in discovering method of web services. After a series of analyses on current algorithm, the paper proposes a matching algorithm used for semantic matching that based on service function. Service classify matching is the first step, service matching can filtrate a number of useless service, leave over the service set that matches user's require. After that is the semantic matching step based on service function, this step needs input parameter-set and output parameter-set for further matching. This paper introduces semantic distance algorithm to service function matching and extends the equal-distance algorithm. At the end of the paper, the author implements the service matching algorithm in an example which proves the feasibility of the matching algorithm.