A Comparative Evaluation of Metaheuristic Optimization Methods for Control Applications

This paper presents a comparative study of metaheuristic techniques for optimizing tuning in two controllers applied to processes characterized by long-dead times. Experimental validation was conducted on an Arduino Temperature Control Lab with additional software-induced delays. The investigation involved tuning the Smith Predictor and PI controllers by utilizing three distinct meta-heuristic optimization algorithms: the Whale Optimization Algorithm, Gray Wolf Optimizer, and Ant Lion Optimizer. The pursuit was guided by the minimization of the Integral Square Error, serving as the cost function. The effectiveness of these control strategies was evaluated using diverse performance indices. The results accentuate the predominance of the Smith Predictor coupled with the Whale Optimization Algorithm, emerging as the most suitable and balanced choice among the examined control methodologies.