图论在产生式系统知识库维护中的应用

当前的专家系统多用产生式规则来表示知识，但是，知识的不一致性和冗余性往往使得专家系统的维护较为困难。本文针对产生式知识库中知识的不一致性和冗余性，提出了一种利用图论来消除这种不一致性和冗余性的方法，并给出了相应的检查算法。该方法思路直观、易于实现，对于产生式系统知识库的维护具有一定的实际意义。

Graph Theory Applied in Maintenance of Rule-based Knowledge Systems

Today, most expert systems use production rules for knowledge representation. As is often the case, non-consistency and superabundance in the rule-based systems make it difficult to maintain them. This paper introduces a new method based on Graph Theory to solve this problem. It also gives the corresponding algorithms. The new idea, clear to understand and easy to implement, proves helpful for maintenance of rule-based knowledge systems.