Research Of Automatic Composition Based On Formalization For Semantic Web Service

Semantic Web and Web service are two important trends of WWW development, the combination of these two technologies results in another new research topic, which is Semantic Web service. Semantic Web service uses semantic Web markup language to describe the semantics of Web service, and makes computer understand the entity to support automatic discovery, execution and combination. The goal of semantic Web-based automatic service composition is to make the software agent automatically combine and complete the task of the user according to the needs of users of various existing Web service, so that service can be flexible to meet the needs of users, to achieve"on-demand service".One important issue in semantic web service composition researching area is how to describe semantic Web service composition formally and verify the correctness of Web service composition. A formal model of semantic Web service composition can be used to check and verify Web service composition so that the correctness of semantic Web service composition can be guaranteed. Semantic Web service composition involves communication and collaboration of several Web service, these characteristics make the verification of the composition of Web service more difficult. Further more, the Ontology Markup Language (OWL) for Web service as to be a service composition method, which is based on semantic, is weak in modeling, and the assurance of the validity is also weak. So the validity of the model and the property of no deadlocks must be simulated and proved before being implemented formally. Aimed at this problem, this paper studies AI planning and Colored Petri nets how to formalize and verify the automatic semantic Web service composition.A semantic Web service composition modeling method based on Pi calculus is presented. This method uses concurrent operators provided by Pi calculus to model a semantic Web service composition as a composition of a set of concurrent Web service. Then, on the basis of this specification, the correctness of Web service composition process is verified. The main points of this paper are as follows:(1) Based on studying how to use AI planning and Colored Petri nets to formalize and verify the automatic semantic Web service composition, the advantages and disadvantages of verifying semantic Web service composition with Colored Petri nets and AI planning are drawed out.(2) Based on the research of the model and its description language of semantic Web service composition and Pi calculus, an approach based on Pi calculus of formalizing and verifying the automatic semantic Web service composition is presented. This method guarantees the correctness of semantic Web service composition, and the cost which dues to the failure of the Web service execution is decreased.(3) A method based on Pi calculus is brought forward to formalize and validate semantic Web service composition,and then a feedback result is presented(4) In this paper, the verification of semantic Web service composition process does not sit at architecture level, but directly on semantic Web service composition language program code, therefore, it will be more feasible and practical.(5) A methed of semantic Web service composition based on Pi calculus is presented. The problem that atomic layer service does not meet the requirements of the logical sub-service is resolved. Not only the correctness of the composition is proved, but also the correctness of the demand is proved.