The Research And Implementation Of Semantic Web Service Matching Model

The traditional Web Service matching is based on the key words matching. However, the lack of semantic information leads to the low recall and precision of this matching. The ability of semantic expression and discovery of web service has been greatly improved by the introduction of the concept "ontology". OWL-S (Web Ontology Language for Services) has also layed the foundation of the research of Semantic Web Service matching.Both of the two current ways of Semantic Web Service matching—basing on reasoning and basing on similarity computing of ontologies, have some limitations: The former strategy has clearly matching structures and high efficiency but can not differentiate matching degrees on the same results level, while the later describes matching degree of the results and the query conditions accurately and enables user to compare among the results, however, it has bad response time in the worst condition and is hard to realize. The current service matching methods are all restricted to the Input and Output matching, but seldom involve the matching of two other import parameters that describe the service ability—Precondition and Effect.On the foundation of service matching strategy based on reasoning, this thesis has proposed a service matching strategy combining the similarity computing, which decides whether to take a similarity computing or not according to the matched level of reasoning matching. It makes optimizations on the service matching process by building up a pretreatment list of ontology concepts to transfer the computing complexity to the pretreatment phase, and improve the performance and efficiency of the matching phase.In addition, the thesis has also done some research about the Precondition and Effect matching of the service: It uses the SWRL-FOL language to describe the Precondition and Effect expressions of the service, and divides the rule expressions into three layers, defines matching rules for each layer, and computes the similarity layer by layer. These have formed a complete I.O.P.E. matching strategy for Semantic Web Service.At last, this thesis has designed an I.O.P.E. matching model for Semantic Web Service and partly realized it. Domain ontologies and service ontologies for demonstration are built up to illustrate the process of service matching, and concept matching experiments are done in "Space" ontology to test the improved effect of the IO matching algorithm.