An improved genetic Hopfield neural networks based on probability model for solving travelling salesman problem

The existing problems of Hopfield neural networks solving travelling salesman problem are analysed and improved energy function is proposed in this paper. Probablity model is introduced into improved HNNs. Probablity model records the gene information of the best individuals,which can make genetic algorithm search simultaneously in depth and width. An improved genetic hopfield neural networks based on probability model is proposed, which not only reduces the rate of invalid tours, but also avoids random search. Simulation experiments show that it can accelerate the convergent speed and enhance the searching ability.