A GENETIC APPROACH TO THE HOPFIELD NEURAL-NETWORK IN THE OPTIMIZATION PROBLEMS

This paper presents an improvement to the basic Hopfield neural network model used for the optimization problems. The method is based on the genetic algorithm. Since the Hopfield network performs gradient descent processing, it suffers from the local minima of the objective function. Global optimization can be achieved by a cyclic process of the network activation. In this paper, a genetic algorithm has been introduced to drive the generation of the initial network state. Thus, more efficient processing can be provided.