A knowledgeable representation language nfa and its application in expert system

This paper presents a method of knowledge representation -- augmented production rule ( namely NFA language ) which integrates logical reasoning and numerical calculation into a unified form. The concept of associate function is introduced in reasoning and calculating network . It augments the representing capability of production rule . Multiple - value logic is used in the design of associate functions to represent the reasoning in the case of incomplete information. Heuristic search functions are used to improve efficiency of reasaning. Certainty factors are included in the conclusion part of some functions to represent inexact reasoning . The control mechanism of NFA language is also mentioned and a successfully used example is presented. This language was used in expert system for oil-well log interpretation which has been running for two years.