Application of Genetic Algorithm to Control Design

A classical PID controller is designed by applying the GA (Genetic Algorithm) which searches the optimal parameters through three major operators of reproduction, crossover and mutation under the given constraints. The GA could minimize the designer's interference and the whole design process could easily be automated. In contrast with other traditional PID design methods which allows for the system output responses only, the design with the GA can take account of the magnitude or the rate of change of control input together with the output responses, which reflects the more realistic situations. Compared with other PIDs designed by the traditional methods such as Ziegler and analytic, the PID by the GA shows the superior response characteristics to those of others with the least control input energy.