Research On Technology Of Automatic Services Composition Based On Semantic Web

Semantic Web and Web service are the two important trends which are emerging in the World Wide Web (WWW). The confluence of Web services and the Semantic Web gives birth to another new technology - Semantic Web Services. The Semantic Web markup language can be used to describe the semantics of Web services and make them computer-interpretable entities, so as to enable programs or agents to locate, invoke and compose Web services automatically. This dissertation focuses on the Semantic Web services composition and aims to develop a technology which can automatically and flexibly compose various existing services to fulfill users' tasks according to their requirements, thus realize so-called service on demand.This dissertation discusses several key issues in Semantic Web services composition automation, including Semantic Web markup language, Semantic Web service description model and services composition method. The main points of this dissertation are described as follows:(1) It designs a Semantic Web rule markup language--OWLRule+ based on thecurrent W3C Web ontology language--OWL.OWL's semantics is limited to description logic, which cannot capture semantics of arbitrary rules. Therefore, OWLRule+ is designed. Its syntax extends OWL and its semantics is based on CARIN, a representation language combining description logic and Horn rules. It does not only provide a logic layer on top of ontology layer for Semantic Web architecture, but also lays the foundation for extension of OWL-S, the current Semantic Web service model.(2) It proposes a new Semantic Web service model called OWL-S+ by enriching OWL-S with the expression power of users' preferences.OWL-S+ model and the corresponding OWL-S+ ontology in OWLRule+ language are defined. The new model contains formal description of user preferences in addition to description of services. It can support automated services composition better than OWL-S, since user preferences affect how to select services.(3) A method for automating services composition is put forward based on OWL-S+ and Situation Calculus. It is named OWL-SC.Situation Calculus is chosen as the formalism for automating services composition. And it is extended to enable expression and handling of users' preferences. Mapping method from OWL-S+ to Situation Calculus is given.(4) A multi-agent system for automating Web services composition, called AWSC-Agents, is designed and a demo is implemented.The architecture of AWSC-Agents and the planning agent algorithm are designed. The concrete implementation mechanism based on IndiGolog, a Situation Calculus-based language and OAA, a multi-agent development framework, is also discussed. The usability and validity of the OWL-S+ model and OWL-SC method is proven by a demo agent system which can plan conference trips for Internet users.