Study on the Improvement of Fuzzy Modeling using GA

Abstract

In the previous research of fuzzy modeling using GA, encode the parameters of antecedent and consequent to binary strings, combine them to form an individual. Therefore, increasing the number of parameters, the length of chromosome gets longer and the convergence to optimal solution can not be guaranteed according to the genetic operator. And the computational cost is high and the performance index could not reflect the correctness of model exactly. In this paper, we propose the identification algorithm of antecedent and consequent parameters using GA and verify the effectiveness through the comparison experiment with the previous method.