基于神经网络的增量式模型算法控制

INTEGRAL MODEL ALGORITHMIC CONTROL BASED ON NEURAL NET COMPUTING

本文把增量式模型算法控制的指标函数映射到Hopfield神经网络的能量函数上,利用神经网络的收敛特性获取控制律,实现了传统控制算法的神经网络化。文中给出了计算机模拟实例。

Projectivity from the performance index of the Integral Model Algorithmic Control onto the energy function of Hopfield neural net is constituted. The control law is obtained by using the property of convergence of neural net and the classic Integral Model Algorithmic Control is performed by neural net computing. The simulation examples demonstrate the effectivness of the algorithm proposed.