关于神经网络的汽车故障诊断专家系统

Expert System for Diagnosis of Automobile Trouble Based on Neural Network

介绍了一种把神经网络与传统的专家系统结合的全集成型的诊断系统．在规则库的建立中，提供了自动的正确性检查及完备性的保证．系统的推理机制是由正向神经网络推理和反向的逻辑推理构成的混合推理系统．推理的完成是采用共同的数据结构和知识表示，以合作的方式来实现的．系统在Ｗｉｎｄｏｗｓ环境下，使用面向对象程序设计实现，为用户提供较方便的接口．

This paper introduces an integrated diagnostic system that combined a neuralnetwork with traditional expert system. In the construction, a series of automaticall method to check correctness and completeness was provided. The inference engine is constituted of two modules. One is the forward neural network inference module; another is backward logical inference module. Inference was completed by cooperate with each other common data structure and knowledge express. The system is constructed under Windows 3 2 Chinese Version Operating System environment, and OOP programming technique are applied. It strived for presenting a convenient interface to users.