A class of PSO-tuned controllers in Lorenz chaotic system

• Published in: Mathematics and computers in simulation Vol. 204; pp. 430 - 449
• Main Authors: Dali, Ali, Abdelmalek, Samir, Bakdi, Azzeddine, Bettayeb, Maamar
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.02.2023
• Subjects: Chaos; Lorenz systems; PSO algorithm; Robustness; Tracking control; Chaos; Robustness; Tracking control; Lorenz systems; PSO algorithm
• ISSN: 0378-4754; 1872-7166; 1872-7166
• DOI: 10.1016/j.matcom.2022.08.016

This paper considers the optimal robust control of chaotic systems subject to parameter perturbations and measurement noise. Three novel dynamic tracking controllers are designed herein to suppress the chaotic behaviour in Lorenz systems under practical constraints. The dynamic controllers allow for extra improvements towards optimal and robust tracking but their design is challenging in the chaotic system. Given a single measured state only, the single-state feedback controllers are optimally tuned using high-performance heuristic Particle Swarm Optimization (PSO) algorithm with a constrained multi-objective function focusing mainly on the Integral Absolute Error (IAE). The controlled system stability is ensured through Routh–Hurwitz stability criteria where control parameters are further tuned using a metaheuristic PSO algorithm to ensure faster transients with minimum errors when tracking desired state trajectories and also to ensure robustness to parameter variations and external disturbances in the chaotic system. Different designs are tested through extensive simulations and comparisons where the proposed PSO-tuned PD-tracking controller is proved superior for not only suppressing the chaotic behaviour and reducing the transients but also for their high robustness to parameter uncertainties and external disturbances. •Novel optimal robust controllers are introduced to suppress the chaotic behaviour in Lorenz systems.•The dynamic controllers ensure system stability and fast convergence under practical conditions.•The single-state feedback controllers are optimally tuned using particle swarm optimization.•High robustness to parameter variations and external disturbances in the chaotic system is attained.•The designs are validated through three scenarios and they have shown superior performance.