Research On Fuzzy Inference Control System's Parameter Optimization Based On GA

Fuzzy controller is widely used in many fields, such as industrial control for its good performance in many complex control problems. However, fuzzy controller is based on reasoning system and the membership functions in this system are mainly denpent on empirical knowledge of experts. These parameters will directly affect the performance of the control. So, it's senstive to make research on the parameters optimization in fuzzy reasoning system-based controller.We optimized the parameters of membership functions of premise and reasoning conclusion using GA, and inproved the method of Homaifar. The method in this paper is based on triangle membership regular fuzzy division which can ensure the consistency and completeness of the reasoning system.For the optimized m-fuzzy division range of control parameter, the dimension of search space is m. For Multivariate problems,our method will much faster than the algorithm proposed by Homaifar.Meanwhile, in the design of genetic algorithm, this individual coding, initial population distribution and selection, crossover and mutation operators make improvements. Through the initial design of the new population distribution and the genetic algorithm mutation operator has the breadth of early assurance that the significant improvement in the latter part of the convergence and improve the general performance of the genetic algorithm.In this paper, the intelligent urban traffic signal control is used as an application example of the proposed application of the method. Optimization of control parameters, the proposed rule variable optimization method is applied to intelligent control of traffic lights into post, by computer simulation to test the proposed algorithm effectiveness.