基于免疫遗传算法的单神经元PID控制器设计

Design of Single-neuron PID Controller Based on Immune Genetic Algorithm

传统单神经元PID控制器中神经元比例系数及学习速率的初始值设置往往依靠经验,或先采用试凑法再根据曲线响应加以微调,针对这一缺点,提出了一种基于免疫遗传算法的单神经元PID控制器。将比例系数及学习速率作为待优化参数,通过参数寻优自行在搜索空间内获得全局最优点。仿真结果表明,新方法自适应能力更强、鲁棒性更好,不仅能够保证较好的控制效果,而且解决了传统方法初始值设置这一大难题。

A single-neuron PID controller based on immune genetic algorithm is presented in the paper to solve initial value set of the proportional coefficient and the learning rate which often rely on the experience, or try at the first and then make a fine-tuning according to the response curve in the traditional single neuron PID controller. The new method put the proportional coefficient and the learning rate as the parameters to be optimized, which gets the global advantage in the search space through the parameters optimization by itself. The simulation results show that the new method has more adaptive capacity and more robustness. It not only can ensure a better control effect, but also solve a big difficult problem of initial value set in the traditional method.