A synthesis method of the approximate reasoning engine by means of genetic algorithm-neural net realization of any multiple-valued logic function using GA

Abstract

In a series of papers, the author proposed an approximate reasoning method which expresses reasoning rules with one newly-defined infinitely-valued threshold function to use it as a reasoning engine, and discussed the advantages and limitations to the fuzzy reasoning. The subject of this paper is the remained problem: how to express the complicated reasoning rules containing non-linearity, non-unateness, etc. The problem is related to the multi-stage synthesis of multiple-valued threshold functions. A synthesis method using the genetic algorithm is devised here with some promising results of realization of arbitrary multiple-valued logic function by threshold functions.