A multiobjective fuzzy PI controller based on genetic algorithm

Abstract

This paper presents a fuzzy proportional-integral (PI) controller, which is a discrete-time version of a conventional PI controller. The data base as well as the constant PI control gains are optimized by using a genetic algorithm according to the following design specifications: minimizing the overshoot, the settling time and smoothing of the output curve. Thus, the optimization problem close to a multiobjective optimization one, resulting in an optimal fuzzy PI controller. Simulation results are shown to demonstrate the improvement over a conventional one.