改进遗传算法优化设计模糊控制系统的研究

Improved Genetic Algorithm for the Research of Fuzzy Control System Design

对T-S模糊控制器的自动设计进行了研究,提出了一种基于遗传算法的自动设计方法。运用并行分布补偿(PDC)思想将T-S模糊控制器分解为多个局部线性控制器,运用极点配置法设计各个局部线性控制器,所配置的极点参数由改进遗传算法进行鲁棒优化。以单极倒立摆为控制对象,运用所提出的方法进行了T-S模糊控制器的设计,实现了对单极倒立摆的模糊控制。仿真结果表明,该方法实现了T-S模糊控制的自动设计,在T-S模糊控制系统的训练精度、鲁棒性等方面具有更好的效果。

To study the automated design of T - S fuzzy controller, a design method of T - S fuzzy controller based on improved genetic algorithm has been developed. Using parallel distributed compensation （PDC） theory, the T-S fuzzy controller is decomposed into several local linear controllers, and each local linear controller is designed by pole assignment method. The pole parameters are configured for robust optimization by improved genetic algorithm. At last, this paper takes the single inverted pendulum as control object and operated it according to T - S fuzzy controller designso as to verify the validity of the design method of T - S fuzzy control system proposed. The simulation results indicate that the improved method makes the automatic design of T - S fuzzy controller possible, and has a better performance in training precision androbust optimization in terms of the T-S fuzzy control system.