Nonlinear system identification based on evolutionary fuzzy modeling

Abstract

The local modeling such as TSK fuzzy modeling is well known as a practical approach for nonlinear system modeling. In this approach, a selection of membership functions makes much effect upon the model performance. It is usually determined by the expert's knowledge for the objective systems. However, it is often difficult to give appropriate membership functions for unknown complex dynamical system without any prior information. In this paper, we deal with the approach to give appropriate fuzzy membership functions based on the observed input and output data using genetic algorithm. Then, an application to identification of nonlinear systems is considered and the availability of the proposed method is illustrated by some numerical examples.