Optimal Constrained Tuning of PI-Controllers via a New PSO-Based Technique
Yuriy Romasevych, Viatcheslav Loveikin, Valeriy Makarets
Abstract
A modified method of particle swarm optimization (PSO) has been developed in the article. It is based on a diversity mechanism (DM), which greatly improves optimization method efficiency. In order to show the enhanced PSO modification, a comparative analysis was conducted. It is based on two proposed criteria and time of the algorithms' execution. The modified method was applied to the problem of constrained optimization of PI-controller tuning. A new criterion was developed with pit-in-pit topology, which includes stability requirement, a set of inequalities (constrains), and a set of optimization criteria to minimize. It allows reducing the initial problem to the problem of unconstrained optimization. The efficiency of the approach is confirmed with four problems of PI-controller tuning.