Research and Development of Software Exception Analysis for Expert System

For the purpose of software testing, it is necessary to employ technologies of software exception analysis. However, software testing needs to be further improved with the occurrence of web techniques. As an assistance of software testing, the software exception analysis of expert system can not only be helpful for the completion of software testing but also guarantee the reliability of testing results.Based on the study of software testing and expert system core technologies (inference method and mechanism, explanation mechanism, knowledge acquirement, and knowledge representation), the author developed the software exception analysis of expert system using Visual Prolog.The expert system diagnoses common exceptions of software running on the Internet, presents exception analysis, puts forward improvement suggestions, explains the analysis process, and provides inference steps. Besides the capability of self-study, the system can continuously enrich the knowledge base with new expert knowledge and enhance its diagnosis and inference power.The system’s desktop-application version runs on Windows with friendly GUI. Itsnetwork-application version, which has B/S three-level architecture, runs on the Internet making use of browser. The expert system and knowledge base are stored on the logic application server in the form of application and on the database server using relational database, respectively.Especially, this thesis explains the overall architecture of the system, and the development techniques of modules like knowledge representation, inference mechanism and knowledge acquirement. It also illustrates the methods of the GUI development with desktop-application version using Visual Prolog. The conversion between knowledge base and relational database for network-application version is discussed as well.