PID Controller Optimization Based on the Self-Organization Genetic Algorithm with Cyclic Mutation

Abstract

This paper proposed a self-organization genetic algorithm with cyclic mutation (SOGACM) and used it to optimize PID controller parameters. A dominant selection operator and a cyclic mutation strategy were given firstly. The former enhances the action of the dominant individuals in the evolutionary process. And the later changes mutation probability periodically in accordance with evolution generation and the period. Moreover mutation probability keeps smaller and crossover operator plays a dominant role in a relatively long period of time. At certain particular time, the probability of mutation increases quickly. The SOGACM was then constructed based on the two operators mentioned above. The analysis of algorithm performance shows the self-organization genetic algorithm with cyclic mutation possesses self-organization property, and has a good global search performance. The simulation results of PID controller optimization experiment indicate that a suitable set of PID parameters could be calculated by SOGACM optimization method.