Multiobjective Optimization for Tuning a PID Controller

The tuning of PI and PID controllers is typically performed by classical techniques such as Ziegler-Nichols and Cohen-Coon, these techniques require interventions in the plants and do not provide suggestions of preference for the responses of the systems.This paper proposes a tuning for a controller applied to a thermal system using a Multiobjective Genetic Algorithm, NSGA-II, and selecting the best tuning through a Multicriteria Decision Making method, the Weighted Sum Aggregated Product Assessment.The results were compared with classic techniques and differential evolution, where one can observe the effectiveness of the proposed method.