A Fact-Based Approach to Define Static and Dynamic Knowledge in Expert System

In this paper, we describe a method for expressing knowledge in expert systems and provide solutions to knowledge representation and internal coupling of knowledge. The term “fact” discussed in the background includes ordinary fact that depicts static system features, and composite fact that depicts dynamic features. The ordinary and composite facts constitute the fundamental elements of expert system production rule. The complex fact is capable of simulating temporal logic semantic, that it enhances expert system knowledge base by adding extensible patterns, therefore, it provides effective tools to extend expert system functionality. Meanwhile, we give an efficient pattern matching method for the inference engine based on this knowledge representation. When dealing with a target system with high complexity and frequent state changes, in comparison to the traditional expert system solution, the fact-based expert system features a simpler infrastructure and more efficient inference pattern matching. In practice, the real-time fault diagnosis expert system we implemented using this method has the basic capabilities of the model checker, and can automatically and effectively match fault patterns in the system state space.