启发式技术在园区网络性能优化中的应用

The Application of Heuristic Technology to Optimize Network Performance

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．

A method to optimize network combination is presented based on heuristic technique. The fitness function is expressed with the traffic between any pair of nodes. The network segment configuration is automatically divided by getting optimal or near optimal solutions in reasonable time. Three methods, clustering, genetic algorithm (GA) and simulation annealing (SA), have been applied to the network segmentation. Among them, SA is superior to GA, and the clustering is inferior under certain conditions.