Genetic Algorithm is a proper method for optimizing Fuzzy controller. This dissertation analyzes the single optimized way of subjection function and control rule, synchronous optimized way and CPM method that Combined Pittsburgh and Michigan approach, and then bring out a new GA for optimizing FLC. It can optimize parameters and structure of FLC simultaneously by doing some betterments for classical GA using a improved selection method, a constrained crossover and mutation operator, self-adapted probability of crossover and mutation, shortening rules of FLC. The results of simulations show that the optimized goal can be achieved based on ensuring the control quality.
|