基于混合优化算法的神经元PID控制策略

Control Strategy for Neuron PID Controller Based on Hybrid Optimization Algorithm

为解决传统比例-积分-微分(PID)控制器在实际工业过程中难以满足控制要求的问题,将二次型性能指标引入到神经元的加权系数的调整中,并利用自学习功能构成了神经元自适应PID控制器。利用混沌优化算法和最速下降法结合起来的混合优化算法,对神经元自适应PID控制器的学习速率和神经元比例系数进行了优化。仿真实验和结果分析表明:该混合优化神经元自适应PID控制器具有很好的动态和静态性能,系统的稳定性和鲁棒性增强,学习参数选择的盲目性和对经验的高度依赖性降低。

Because it is difficult for the conventional Proportional Integral Differential（PID） controller to meet the control requirements in the actual industrial control process, a self-adaptive PID controller of neural network is designed in this paper. It imports the quadratic-form performance index to the setting of neuron weight coefficient and utilizes the self-learning function of neuron. Furthermore, hybrid optimization algorithm combining the chaos optimization algorithm and the steepest descent method is used to search for the optimum parameters of the self-adaptive PID controller of neural network. Simulation experimental results and experimental analysis prove the superiority of the hybrid optimum neural PID controller. It has better dynamical and static performance, the fitness and robustness of the system are strengthened, and the blindness of selecting learning factors and the high dependency on experience are debased.