Nature-Inspired Metaheuristics for Tuning the PI Controller of a High-Voltage Pulse Generator

Abstract

Metaheuristics are currently playing an increasingly important role in the tuning of industrial controllers. In particular, this paper presents the results of implementing various nature-inspired metaheuristics for the tuning of a proportional–integral controller intended for use in a high-voltage pulse generator. This paper analyses and compares the results of tuning obtained using both classical metaheuristics, such as simulated annealing, genetic algorithms, particle swarm optimisation, and differential evolution, and newer approaches, such as sand cat swarm optimisation and sea lion optimisation. An original, complex multi-criteria cost function is constructed in this paper for optimising and ranking nature-inspired metaheuristics for the tuning of the proportional-integral controller. The results show that sand cat swarm optimisation outperforms other optimisation approaches according to the adopted multi-criteria optimisation criterion.