遗传算法改进的单神经元PID控制器及其应用

The Application of Single-neuron PID Improved by Genetic Algorithms

单神经元(PSD)控制器利用神经元的自学习、自组织能力,根据被控对象的变化情况对控制器的权值进行在线调整,达到了在线调整PID参数的目的;且设计时无需精确对象模型,克服了常规PID因负载、模型参数的变化及非线性因素等影响PID精确调节的缺点。采用遗传算法对单神经元PID参数进行寻优,在搜索空间内获得全局最优点。仿真结果表明:采用遗传算法改进的单神经元PID控制器能获得较好的控制效果。

Single-neuron PSD Controller has the ability of on-line change of weights to reach the purpose of regulating the parameters of the PID by using the neuron’s self-studying and self-organization and the change of the controlled plant, which overcomes the disadvantages of affecting a conventional PID’s accurate regulation because of the change of load, model. A genetic algorithm is used to search for the optimal PID parameters. The simulation shows it can obtain a good result.