A fuzzy-genetic approach for automatic tuning of a PID controller

A novel method employing fuzzy logic and a genetic algorithm for automatic tuning of a Proportional Integral Derivative (PID) controller is presented. The technique adopted is based on the theory of dead-beat control. A fuzzy logic technique has been used to predict the controller output and the crisp consequent values of the rulebase on the Takagi-Sugeno model are optimized using a genetic algorithm. The proposition is an extension of the work by R. Bandyopadhyay and D. Patranabis (2001), where the rulebase was prepared based on the knowledge of process experts. Significant improvement has been obtained using a genetic algorithm by optimizing the crisp consequent values of the rulebase. As can be seen from the simulated results, the method shows substantial improvement over the controller tuned with the Ziegler-Nichols formula (J.G. Ziegler and N.B. Nichols, 1942) and the PID controller proposed by Bandyopadhyay and Patranabis.