The Study On Fuzzy Control System By Genetic Algorithms

Fuzzy controllers perform very well in some difficult control problems. But it is very difficult to confirm the parameters of a fuzzy controller. Usually the membership functions and fuzzy rules are defined on the base of workers' experiments and experts' knowledge. In this paper, the ability of optimization of Genetic Algorithms is used to encode and optimize the parameters of membership functions and the consequents of fuzzy rules.Because of the large time lap of temperature systems, general controllers perform badly in such systems. So in this paper, the temperature of a house is made as the object of this fuzzy controller to check up the performance of the fuzzy controller. First, the software of MATLAB is used to program the Genetic Algorithms. The parameters of membership functions and the consequents of fuzzy rules are taken as the optimized answers. After the operation of encoding, selecting, crossover and mutation, the optimum parameters of these membership functions and the optimum consequences of these fuzzy rules are confirmed. So the design of this fuzzy controller is accomplished. Then, the math model of this fuzzy controller system is established, using the software of MATLAB. From the result of calefactive and temperature-fall processing, the good performance of this fuzzy controller is confirmed. In the last, the hardware of this fuzzy controller is designed, and the flow chart of looking-up table software is given.