A Knowledge Representation Method for Developing an Expert System : Page Model

This paper proposes a knowledge representation method (page model) for developing an expert system. This method has three characteristics : (1) It classifies the knowledge into facts, relation-ships among facts, rules, and an inference mechanism which is a certain arrangement of rules for inference. (2) It notes the features of each knowledge component in one page in such a manner that a theory is usually evolved by using notes. It distinguishes between rules which have a theoretical background and empirical rules for a particular problem. (3) It proposes six symbols to represent the inference mechanism : AND convergence, OR convergence, AND branch, OR branch, backtracking and cut. This model is applied to represent an expert system for a multi-item assembly line, and its advantages are evident.