遗传算法在PID自整定控制中的应用

Application of Genetic Algorithm in Self-tuning PID Controller

提出了一种基于遗传算法和单神经元的自整定PID控制器的设计方法,该控制器首先利用遗传算法对PID的3个参数作离线优化,搜索到一组准最优的PID参数,作为PID控制器参数的初始值,然后利用改进后的单神经元梯度下降法在线调节PID参数,以使系统获得最优的动态性能和稳态性能.仿真结果表明:与传统PID控制算法比较,该控制方法响应速度快,具有更好的控制效果.

In this paper, the design of a self-tuning PID controller based on genetic algorithm and single nerual cell is presented. By means of the off-line optimization, based on genetic algorithm, of three parameters, a group of optimal PID parameters is obtained, which is further used as the initial input parameters of PID controller. Based on these initial PID parameters, through the improved gradinent decent method of single neural cell, all PID parameters are adjusted on-line to ensure the system with optimal dynamic and steady performance. Simulation results indicate that, comparing with the traditional PID controller, this one may increase the respond speed and lead to more effective control.