Adaptive Constrained Predictive PID Controller via Particle Swarm Optimization

As an alternative to genetic algorithm, particle swarm optimization (PSO) is a new population-based evolutionary technique and has been attracting much attention to apply in different fields, such as nonlinear programming problems and neural network training. In this paper, a novel time-varying adaptive constrained predictive PID controller via PSO is proposed. This is based on the optimization of the GPC criterion with considering the constraints on the parameters of PID structures and control signal. Furthermore, PSO and non-differentiable exact penalty function technique are utilized to obtain the adaptive constrained predictive PID controller parameters. The proposed controller is suitable for different order systems and does not require the control horizon to be equal to one. As PSO is robust under the presence of nonlinear structures in the performance index and constraints, the proposed controlled can be easily applied to different problems. The simulation results show that the proposed controller is effective.