The Study of Knowledge Representation and Uncertainty Reasoning for Computer Aided Failure Analysis Expert System

摘要

Knowledge acquisition and representation of failure analysis as well as the reliability of conclusion are studied in this paper. The ways of knowledge acquisition and representation are given through the comparison of various knowledge acquisition algorithms and representation methods. The document-based processing algorithm is mainly used to acquire domain knowledge. The knowledge of failure analysis can be classified as the basic knowledge and case knowledge. Rules aided by procedural and object-oriented scheme are appropriate for representing the basic knowledge, and frame method for the failure case knowledge. The classical probability-based method is suitable for describing the strength of backward chaining rules, and certainty factor for describing the strength of forward chaining rules. As for the failure case knowledge, it is considered that case adaptation can incorporate the uncertainty reasoning.