Research On Method Of Web Services Composition Oriented To Credit Evaluation

Nowadays, there are already many Web services are running on the Internet, and the number of service is still increase rapidly, but the vast majority of these services are small size relatively, and single function, so that this form is difficult to meet the growing and changing needs of users. So, the problem that how to composite these existing Web services which are running on heterogeneous platform in an appropriate way, for providing users a composition services with functional requirements and performance requirements, has become a hot spot research field.The research on Web service composition involves numerous problems that often exist in establishing time and runtime of the service composition. The former includes service description, matching and selection, auto-composition and verification, etc., and the latter includes service composition monitoring and executing, security and transaction management, etc. This paper mainly researches the issues in the process of establishing service composition, including how to describe requirements of service composition, how to match and select service, and how to composite service automatically.At present, a lot of Web service composition methods have been proposed. These methods mainly divided into two types: workflow-based Web service composition and AI Planning- based Web service composition. However, there are some problems in workflow-based service composition, such as low efficiency automation process and insufficient flexibility; and some time the method of AI-based service composition is too high complexity to achieve.Based on the above, the mainly research contents of this paper include the following aspects: 1) According to the composition requirements in credit evaluation domain, in this paper, we use Ontology Feature Model (FM) to describe the features and their dependencies. Service FMs can provide semantic support for matching and selection, their dependencies can help establish composition process and test combination among services.2) In order to get the most service to meet the needs of composition, this paper, taking semantic matching and QoS-based selection into account, uses hybrid selection strategy and mainly focuses on four types of control structure in service composition.3) Focuses on the problem that no single concrete service to meet the needs of abstract service node, the algorithm of back searching Semantic Sequence Matrix is proposed, it can composite Web service automatically for the target needs.4) Finally, this paper designs and implements the service composition instance of credit rating based on the ontology feature model and the abstract process of service composition.