Determination of the gains used in model reference learning control method with a genetic algorithm

Abstract

Fuzzy Model Reference Learning Control (FMRLC) is a method which provides a systematic design procedure for the design of fuzzy controllers. In most classical fuzzy control systems, some parameters of the membership functions are determined by trial and error method. However, in FMRLC method, these parameters are determined by a learning mechanism. In this paper, the FMRLC method is investigated and applied to the control for the cargo ship steering and, simulation results are presented. During the investigations, it is found that the input and output gains used in the FMRLC system are very effective on the control performance and there is no systematic way in the literature to determine the gains. Therefore, in this study, a Genetic Algorithms (GA) is used to find the input gains which will provide a good control performance. Simulation results showing the effectiveness of the proposed control structure are given in the paper.