Critical analysis of hopfield's neural network model for TSP and its comparison with heuristic algorithm for shortest path computation

For shortest path computation, Travelling-Salesman problem is NP-complete and is among the intensively studied optimization problems. Hopfield and Tank's proposed neural network based approach, for solving TSP, is discussed. Since original Hopfield's model suffers from some limitations as the number of cities increase, some modifications are discussed for better performance. With the increase in the number of cities, the best solutions provided by original Hopfield's neural network were considered to be far away from those provided by Lin and Kernighan using Heuristic algorithm. Results of both approaches are compared for different number of cities and are analyzed properly.