提出了一种采用启发式技术用于网络组合优化的思想．通过确定网络适合度函数，在合理时间内求解最优解或准最优解，能自动地对网络进行合理分割．通过模拟聚类、遗传算法（ＧＡ）、模拟退火（ＳＡ）等三种方法在网络划分中的应用，在一定条件下，ＳＡ的性能最优，ＧＡ次之．研究结果表明，用此方法是行之有效的． Proposed a heuristic technology for network combination optimization ideas. By determining the fitness function of the network, the optimal or quasi-optimal solution can be solved within a reasonable time and the network can be reasonably partitioned. Through the application of three methods of simulation clustering, genetic algorithm (GA) and simulated annealing (SA) in network partitioning, under certain conditions, the performance of SA is the best, followed by GA. The results show that this method is effective.