In this paper, I propose the improved algorithm for the membership function modification, which can improve the performance of the Fuzzy Logic Controller (FLC) by representing the control knowledge with the experts and the operators more exactly. This...
In this paper, I propose the improved algorithm for the membership function modification, which can improve the performance of the Fuzzy Logic Controller (FLC) by representing the control knowledge with the experts and the operators more exactly. This paper suggests an algorithm to modify the position and shape of the fuzzy membership function based on the input-output data clustering so that the FLC can represent the control knowledge more exactly. It is used by the rough control knowledge retrieved from the intuitive experiential knowledge and experience as the evaluation criteria to cluster the input-output data.
I applied this algorithm to the model for the traffic signal, and showed that it can improve the performance of the FLC so that it can solve the difficulty of the range partition of the linguistic variables.