A design method of optimal control system by use of neural network

This paper proposes a method to design an optimal control system by use of a neural network. Because the normal backpropagation (BP) method cannot be applied to this case, we choose Powell's conjugate direction algorithm for training the neural network. In a regulator problem, the neural network functions as a state feedback controller, and in a servo problem it functions as both feedfoward and feedback controller. The proposed method can be applied to various problem where conventional methods cannot be applied. Simulation results show the effectivity of the proposed method.