Optimization of PID parameters using Genetic Algorithm and Particle Swarm Optimization

Optimum PID controller design using real coded genetic algorithm (RGA) and particle swarm optimization (PSO) is addressed in this paper. The lower and upper bound of the design variables (K_P, T_I and T_D) are selected in an intelligent manner around the values obtained from Ziegler-Nichols method. PID controller was designed by minimizing various performance measures such as ISE, IAE and ITAE for two different linear systems. The performance of RGA and PSO is compared with respect to time-response specifications, computation time and statistical performance in 20 independent trials. The simulation reveals that the performance of PSO and RGA with simulated binary crossover (SBX) produced almost same time-domain performances and take same computation time. Also, it is found that PSO converges with less number of functional evaluations, consistent for simple systems and in general ITAE is preferable for quick settling time.