Hopfield network and genetic algorithm to solve n-p hard optimization problem

Abstract EN

This paper presents a hybrid Hopfield neural network (HNN)-genetic algorithm (GA) approach to tackle the combinatorial optimization problem.

GAs is global search algorithms.

On the other hand, in HNN the range of a search is neighborhood of the initial point.

Then HNN is local search algorithm.

Our hybrid system utilizes the features of these algorithms that make up for defects of each other, so we can overcome some difficulties, such as coding and crossover in GAs and setting up the initial point and parameters in HNN model.

The effectiveness of our approach is demonstrated through reinvestigation of its capability of solving the traveling.