专家系统中多种知识表示方法的集成应用

On Integration of Several Knowledge Representations in the Expert System

专家系统的建立是以知识为基础的,采用何种知识表示方法在一定程度上决定了专家系统能否正确高效的运行。先综合比较了常见的几种知识表示方法,然后介绍了在“现役混凝土结构耐久性评估专家系统”1中集成应用框架、产生式、神经网络等多种知识表示方法的思路及解决方案,该系统的实际使用证明此法是可行和有效的。

The efficiency and effectiveness of an Expert system depends somehow on its knowledge representation. This article first makes a comparison among several knowledge representations used in expert systems, and then proposes an integrated solution of multiple knowledge representations such as frame, production and neural network. It is illustrated with an example of Expert System for Durability Assessment of Reinforced Concrete Structures, which proves effective and efficient.