Identification of Takagi-Sugeno (TS) fuzzy model with Evolutionary Parallel Gradient Search

In this paper the modeling of nonlinear system with TS fuzzy model is discussed. The identification of TS fuzzy model is first posed as an optimization problem and a new hybrid optimization algorithm- referred to as evolutionary parallel gradient search (EPGS) is applied to find the optimal values of the parameters in the fuzzy model. The main feature of EPGS is its ability to deal with the local minima problem in global optimization. By using the gradient information of cost function, EPGS combines gradient-based algorithm and evolutionary algorithm (EA) in an innovative way such that EA is used to keep the best searches at every step in the optimization process and the gradient descent method is used to update these best searches. The application of EPGS in the parameter estimation problem of TS fuzzy models shows excellent performance in terms of modeling accuracy.