Study on optimization of fuzzy controller based on genetic algorithm of population classification

This article referred to the thought of species migrating mechanism to improve the genetic algorithm, the population was divided into sub-populations of three levels, as follows: low-level population, sub-optimal population, senior population. The fitness of three populations and the encoding digits of individual gene chain increased gradually, therefore, the possibility of producing higher fitness individuals would improve greatly, and could enhance the search capability of genetic algorithm, achieving the global optimum while guarantee local optimum. And parameters of fuzzy control to be optimized were reduced by using shrinking factors. Membership functions of fuzzy control, scaling factor and quantitative factors were optimized by the improved genetic algorithm. Finally, for a second-order time delay system to simulate, the result shows that the quality of control of fuzzy controller optimized acquires greater improvement and enhancement.