【摘要】 本文应用Hopfield神经网络模拟方法对二部图进行判别。提出了邻域矩阵的概念。将任意的连通图输入至神经网络判别系统,输出该图的邻域矩阵,对应于系统能量函数取最小值的输出状态即为二部图邻域矩阵,同时得到该二部图的顶点划分;如能量函数非最小值,则判定该图不是二部图。该判别法的核心是构造一种广义的能量函数——Liapunov函数,使原来难以解决的问题找到新的解决途径。更多还原

【Abstract】 This paper uses Hopfield neural network modelling metbod to distinguish dual-graph, building up the concept of neighborhcod matrix. Input a random connected graph to the network, output its neighborhood matrix, when the minimum value of energy function can be taken, the output is a dual-graph neighborhocd matrix, while the vertex division is acquired; if not, it is not a dual-gragh. The core of the distinguishing method is to construct a general energy function——Liapunov function.更多还原