NEURAL INFORMATION PROCESSING, vol.3316, pp.254-260, 2004 (SCI-Expanded)
In this paper, differential evolution algorithm (DEA), one of the most promising Evolutionary Algorithm's, was employed to tune a PID controller and to design a set-point filter for unstable and integrating processes with time delay. The proposed cost function used in DEA gives the shortest trajectory with minimum time in the phase plane. The results obtained from the proposed tuning method here were also compared with the results of the method used in [1]. A time-domain cost function is deployed in order to obtain good compromise between the input step response and disturbance rejection design. The PID controllers optimized with DE algorithm and the proposed cost function gives a performance that is at least as good as that of the PID tuning method from [1]. With PID tuning method using DEA, a faster settling time, less or no overshoot and higher robustness were obtained. Furthermore, the tuning method used here is successful in the presence of high noise.