Research On Web Service Discovery Based On OWL-S

With the rapid development of web service technology recently, the growing number of web services increases dramatically. How to efficiently locate functionality-desired web services among lots of web services and select a best one among large lots of functionality-similar web services becomes the hot topic of research. The current mainstream web service matching techniques stress overmuch on the rigid"matching"between requests and service advertisements. However, in real-world applications, such accurate matching is not always feasible, nor is sometimes desirable for users. To cater for the vagueness of the user requests, there is a need to pursue more flexible matching technologies. To solve the above problems, the paper does work on the key technologies including semantic description model, semantic web language and the matching algorithm.In the description of web services, the web services in the extended manufacture field are described by OWL-S semantic model in this paper. Thus the functions of service are described in the level of semantic.In the web services matching models, the paper designs the semantic web-based service matching layered model based on UDDI, including three tiers. The first tier is called service category matching, which compares the similarity between a request and a service advertisement in terms of the"Service Category"description in the OWL-S Service Profile sub-model. The second tier is called precondition matching, filtering out the services that don't satisfy the precondition. The third tier including functional matching and quality of service matching is the core of the matching model. The matching of the first tier and the second tier decreases the computational load of the third tier matchmaking.The paper presents a hybrid semantic matching algorithm to evaluate the similarity degree between the demand and the description. The hybrid semantic algorithm combines the advantages of the algorithms based on geometry distance and information content, considering the concept of information capacity and the strength of edge. The hybrid semantic algorithm solves the problem that the geometry distance and the information content don't well consider the depth and the density of the Ontology concept.Finally, based on the extended manufacture ontology, a prototype system for semantic matching of web services is developed, the model presented in this paper and the methods are verified.