A multi-population genetic algorithm approach for PID controller auto-tuning

The present paper applies a multi-population genetic algorithm (MPGA) to the Proportional, Integral and Derivative (PID) controller tuning problem. Two control criteria were optimized, the integral of the time multiplied by the absolute error (ITAE), and the integral of the time multiplied by the absolute output (ITAY). The MPGA is compared with a standard genetic algorithms (SGA) already applied to the same control model. The control criteria are supplied by neural networks (NN) previously trained for this purpose. The control tuning and the corresponding responses were obtained using the MATLAB/SIMULNK environment. The computational results show a superior performance of the MPGA even when compared with the exact values found by dynamic simulation using gradient techniques.