A ring-cache genetic algorithm for tuning reliable fuzzy logic controller

We improve the organization of fuzzy logic controller by incorporating two strategies: the ring-Cache GA and reliability analysis. In our previous papers (1995, 1996), we proposed the self-tuning methods all based on GA. However, when the size of the population grows up and the encoding of the chromosome is very large, the expense of the evolutionary computation would be out of our expectation. In this paper, we combine the ring GA and cache GA to derive a ring-cache GA for improving the efficiency of GA. Another important problem exists in the fuzzy logic controller: when we use GA to tune the controller, the system does not promise to hit the target or goal every time. So we need to define a reliable region or fitness bound to make sure the controller is reliable. In this paper, we enhance the reliability analysis with the original fuzzy logic controller to form a reliable genetic fuzzy logic controller. We use our approach to solve the robot path planning problem