Research On Services Composition Supported By Case-Based Reasoning

With rapid development of internet technology and application, Web Services are considered as self-contained, self-describing, modular applications that can be published, located, and invoked across the web. Nowadays, an increasing amount of companies and organizations implement their core business and outsource other applications as Web Services over Internet. However, single Web Service cannot satisfy all the functionality needed by the user. It is imperative task to study how to compose Web Services into a new value added service at low cost instantly.This dissertation studies how to make Web service composition that could satisfy the diversity of users' demand, knowledge expression and accumulation as well. The main contributions are as follows:This paper presents Web service composition frame supported by Case-Based Reasoning after analyzing the characteristics of Web service composition and advantage of Case-Based Reasoning technology. Service cases, which are design by domain experts and are used in Web service composition, can be provided to user by the frame transparently. Through retrieval, revision, reuse and retain mechanism, which are designed properly, Web services composition that satisfy users' demand can be realized with more accuracy and extensibility at low cost.The idea of integration and abstraction of the solution part of the service case is presented to solve problems of case construction, maintenance and reuse during the procedure of Web service composition supported by Case-Based Reasoning. According to this idea, service cases can be composed, revised and reused at high abstract levels, ability to adaptation and reuse can be improved as well, difficulty of service cases' development and revision can be reduced in addition.A new method of service cases construction is presented according to characteristics of Web services representation and Case-Based Reasoning. Furthermore, a unified case oriented representation method, which is based on RDF and abundant in semantic, is provide to be easy to integrated case knowledge and domain knowledge. The integration of case knowledge and general domain knowledge ontologies could make sharing different cases" knowledge among case bases be possible, and realize case representation, construction and retrieval.The service case which is chosen by similarity degree measurement based basic description can neither satisfy all demands of users nor revise easily. To solve this problem, a similarity degree measurement, which is induced into three hierarch models, is presented. Furthermore, similarity functions with rough set theory are designed to solve problems of lack of adaptations, which reduces computing complexion and improves matching efficiency.According to Web services characteristics, a revision method based on hierarchical service case is presented to improve flexibility and extensibility of the method of web services composition based on Case-Based Reasoning. Moreover, on the basis of abstraction, a method of revising service case adaptively, which is based on the procedure of operator revising the most similar service case chosen from case base and based on interaction environment modeling, is presented. According to this method, on the basis of ontology, the requirement of dynamic revising service cases and the description of revision operator are presented under the same circumstance, and the capacity of revision operator can be described with two respects of interaction environment and changes. Furthermore, the model of resource states transition graph is presented to realize the transition from the description of the capacity of revision operator to the resource states of revision operator. Finally, a verification method concerning the dynamic revision of service cases is presented to improve the revision and verification automatically.