本文发现了货郎问题(TSP)能量函数的参数空间中的吸引域与全局性优化解之间存在着确定的对应关系,并找到了许多参数空间中的吸引域。在此基础上提出了一种新算法——参数正交选优神经网络算法求解TSP,30城市,31城市(中国)和300城市TSP的求解结果都十分满意,表明这种新算法的有效性。

<正> For solving the map-coloring problems,this paper presents an energy function,amore effective dynamic equation and a more simple convergence condition.For the first time westudy the map-coloring problems in the way of connecting discrete Hopfield neural network withthe orthogonal optimization,and as a practical example,a color map of China is given.