Identification of dynamic systems using a genetic algorithm-based fuzzy wavelet neural network approach

Abstract

In this work, a genetic algorithm based approach for designing fuzzy wavelet neural network (FWNN) is developed. The FWNN approach combines fuzzy set theory and wavelet neural networks. Thus, the proposed Fuzzy WNN is implemented through an interconnected network including two network structures, one containing the fuzzy reasoning mechanism and the other containing Wavelet neural networks. Then a simple genetic algorithm is used to find optimal values of the parameters of the both network structures. The ability of the technique FWNN in identifying non linear dynamical systems is demonstrated on two examples.