An enhanced inference strategy for machine fault diagnosis system using hill-climbing approach

Most fault diagnosis systems in mechanic domain usually emphasize on the correctness of the hypothesized result. In time constrained situations, the efficiency of the diagnostic process becomes more important and should not be overlooked. This paper proposes a new inference strategy that can enhance the efficiency of the diagnostic process. Specifically, the proposed inference strategy attempts to find out the most efficient diagnostic process for detecting the cause of machine malfunction by the aid of multi-attribute decision making (MADM) method and hill-climbing search method. To verify the performance of the proposed inference strategy, two VCR diagnosis expert systems have been generated. One is developed by applying the proposed approach, and the other one is an ordinary expert system. The evaluation result indicates that the former system requests less diagnosis time than the latter system does when forty real repair records are employed.