Improved Modeling and Optimization of T-S Fuzzy Models

Fuzzy modeling is an effective method for nonlinear systems, but there exist many unsolved issues due to the complexity of nonlinear system. This paper proposes an improved modeling and optimizing method for T-S fuzzy models. Firstly, we combine the fast search method of density peaks with the fuzzy cluster method (FCM), in which the former is utilized to find the initial clustering center and then the latter achieves the cluster. Secondly, the initial T-S fuzzy model is obtained by using the least square method to identify these parameters. And then, an improved differential evolution algorithm is utilized to optimize the above structure and parameters. Finally, the experimental results over a representative example show that the proposed method can improve the identification precision and convergence speed for T-S fuzzy model.