Optimization of Fuzzy Logic Controllers with Rule Base Size Reduction using Genetic Algorithms

In this paper, we present the automatic design methods with rule base size reduction for fuzzy logic controllers (FLCs) through real and binary coded coupled genetic algorithms (GAs). The adaptive schema is divided into two phases: the first phase is concerned with optimizing the FLCs membership functions and second phase called rule learning and reducing phase which automatically generates the fuzzy rules as well as determines the minimum number of rules required for building the fuzzy models. In the second phase, the redundant rules are removed by setting their all consequent weight factor to zero and merging the conflicting rules during the learning process. The first and second phases are carried out by the real and binary coded coupled GAs, respectively. Optimizing the MFs with learning and reducing rule base concurrently represents a way to maximize the performance of a FLC. The control algorithm is successfully tested for intelligent control of two degrees of freedom inverted pendulum. Finally, the simulation studies exhibits the better or competitive performance of the proposed method when compared with the existing methods.