A continuous hopfield neural network based on dynamic step for the traveling salesman problem

Abstract

For the traveling salesman problem (TSP) which is also an important aspect for mobile robots, a continuous Hopfield neural network based on dynamic step is applied to solve TSP. For combinatorial optimization problems such as TSP can be mapped to a Continuous Hopfield neural network (CHNN). The dynamic step size is used to replace the fixed step size, which can solve the problem of the mutual restriction between the convergence precision and the convergence speed. The energy function is designed to represent the path length. The energy of network is constantly updated and converged to a minimum value eventually. Meanwhile, the optimal solution is obtained for TSP. Simulation results show that the proposed algorithm can accelerate the convergence rate and obtain high precision optimization results for TSP.