Modelling Type-2 Fuzzy Systems by Optimized Nonstationary Fuzzy Sets with Genetic Algorithm

Abstract

The high computational complexity of type-2 fuzzy systems causes emerging of alternative methods. Nonstationary fuzzy system, which aims to model the type-2 fuzzy sets with a number of type-1 sets -obtained with the help of perturbation function- is one of these methods. In this study, the type-1 fuzzy subsets which are used in nonstationary systems are optimized by the help of the genetic algorithms. Thanks to the convergence rate of the genetic algorithm, the obtained type-1 fuzzy subsystems are close to the best solution. So, the nonstationary fuzzy set which is generated using the genetic algorithm gives us a better solution instead of the nonstationary fuzzy sets created by perturbation functions which are based on mostly randomness. The success of the obtained nonstationary fuzz set is proven by the simulation results.