A New Approach for the Design and Implementation of Fuzzy Controllers
Abdollah Homaifar, E. McCormick · 2005
Genetic algorithms (GAs) are powerful search procedures based on the mechanics of natural selection, Using operations found in natural genetics to guide them, they provide a means to search poorly understood, irregular spaces. Fuzzy systems arose from the desire to describe complex systems with simple tools. Where membership in a set of a boolean system is either {1} or {0}, fuzzy systems allow for degrees of membership over the range {0-1}. This imitates the linguistic approach to describing conditions (i.e. cold, very warm) used in everyday life. In this manner, fuzzy theory can form controllers using rules of the form lF{condition} THEN{action}. Previous work on this subject has focused on the design of either membership functions or the design of rule sets. In this paper we examine the applicability of using GAs in the simultaneous design of rule sets and membership functions.