An iterative genetic learning approach for Takagi-Sugeno fuzzy systems

In this paper, we propose an iterative genetic learning approach for automatically finding an appropriate set of membership functions and rules in Takagi-Sugeno fuzzy systems. The proposed approach can be divided into two parts. In the first part, an initial rule set is given and the set of membership functions appropriate for the TS fuzzy model is automatically induced. In the second part, the induced membership functions are then fed as the input into the TS fuzzy model to obtain a new rule set. The rule set is then fed into the first part to repeat the same learning process. The process is thus executed iteratively until the terminal criterion is satisfied. The experimental results show that the proposed approach can derive good membership functions with good shapes and rule sets with low errors in Takagi-Sugeno systems.