Abstract:
This paper investigates the problem of an optimal control for linear systems with time
delay and its solution by means of Hopfield neural networks. First, we discrete the
system and make it like a system without disturbance then design a disturbance
compensator for the system. Second, the dynamic optimization problem is
transformed into a static optimization problem via linear state space analysis
methods. The parameters of the Hopfield neural network are adjusted such that the
network solves the static quadratic optimization problem yielding the optimal control
sequence. The outputs of the neurons of the network represent the values of the
optimal control signal in each time step.