Controller tuning with multi-objective optimization: benchmark definitions for SISO linear systems: A PI controller case

Abstract

Multi-objective optimization techniques are practical techniques for controller tuning purposes. A wide variety of papers use such an approach or propose new algorithms to approximate the Pareto front. Nevertheless, despite the expressive volume of works dealing with it, there is no standard benchmark testbed that could be used as a baseline comparison among different techniques. This work addresses such a gap, with a first step proposing a linear-SISO testbed with guidelines and rules and providing a basic example for a PI controller. Such a benchmark is expected to promote further research on the topic.