【摘要】 本文应用神经网络研究通信网设计中的划分问题。文中给出了一种基于神经网络求解图划分问题的并行算法；该算法验证了一些随机生成的例子（包括高达１００个顶点的问题）。计算机模拟结果表明该算法在１５步迭代之内求得较好解，优于传统的启发式算法。更多还原

【Abstract】 In this paper, the artificial neural network is proposed for the partitioning problem in telecommunication network design.A parallel algorithm based on the neural network is presented and used to solve the graph partitioning problem. The parallel algorithm was tested by some randomly generated examples including up to 100-vertex problem.Computer simulation results show that the parallel algorithm finds better solutions than the heuristic algorithm within fifteen iterationn steps.更多还原