Параметрична оптимізація нечітких систем керування на основі гібридних ройових алгоритмів з елітною стратегією

• Дата: 2025
• Автори: Kondratenko , Yu.P., Kozlov , A.V.
• Формат: Стаття
• Мова: English
• Опубліковано: V.M. Glushkov Institute of Cybernetics of NAS of Ukraine 2025
• Теми: нечітка система керування; параметрична оптимізація; мультиагентний метод оптимізації; алгоритм рою частинок; нечіткий регулятор; мобільний робот
• Онлайн доступ: https://jais.net.ua/index.php/files/article/view/691
• Назва журналу: Problems of Control and Informatics

This article is dedicated to the development and study of a hybrid multiagent method of parametric optimization of fuzzy control systems that combines the advantages of particle swarm algorithms and local search algorithms based on the elite strategy. The obtained method allows to optimize effectively various parameters of fuzzy systems, finding the global optimum of the problem to be solved, and, at the same time, has higher convergence rate compared with the basic particle swarm method. The effectiveness study of the proposed hybrid multiagent method is carried out at parametric optimization of a fuzzy speed control system for the multipurpose caterpillar mobile robot able to move along inclined and vertical ferromagnetic surfaces. In particular, in this paper, the problem of parametric optimization of the weight coefficients of the consequents of the rule base of the fuzzy speed controller for the mobile robot is solved, using the basic particle swarm method, as well as using two modifications of the developed method: the hybrid particle swarm method based on the elite strategy with gradient descent, the hybrid particle swarm method based on the elite strategy with the algorithm of extended Kalman filter. The analysis of the computer simulation results showed that both modifications of the proposed hybrid method allow to conduct optimization of the coefficients of the consequents of the rule base significantly more efficient compared with the basic particle swarm method. Also, a speed control system with an optimized vector of coefficients of the controller's rule base based on the considered modifications of the hybrid method of parametric optimization has higher quality indicators of control than a system with an optimized vector of coefficients based on the basic particle swarm method. Thus, the studies conducted in this article confirm the high efficiency of the developed hybrid multiagent method based on the elite strategy, as well as the feasibility of its use for parametric optimization of fuzzy control systems of various types and configurations.