Hopfield网络求解TSP两种改进算法的仿真研究

Simulation research of two improved methods of Hopfield network to solve TSP

针对Hopfield神经网络在求解旅行商问题(TSP)时出现的无效解和局部极小值问题,分析和比较两种改进的求解方法,首先从理论上证明算法的有效性,然后对两种算法分别进行计算机仿真,探讨网络收敛于全局有效解的途径。研究表明,改进的算法都可避免无效解,在求解10城市问题时可获得近乎100%的最优解。

Aiming at the difficulties of invalid solution and local minimum problems when traveling salesman problem （TSP）is studied using Hopfield artificial neural network.Firstly two improved algorithms are compared and analyzed;the effectiveness of these algorithms is confirmed theoretically, then these algorithms are simulated separately. Way of conver- gence overall effective solution are diseussed.A large number of simulation results show that the improved algorithms have the advantages of avoiding most of the invalid solution,and obtaining the optimum solution nearly at the percent of 100 when solving 10 city problems.