基于排队论指导的K-Means聚类算法及其在TTC网络优化设计中的应用

提出了一种在K-Means算法基础上改进的聚类算法QSKM,对网络结点分组进行优化设计。我们利用排队论对网络通信中的呼叫与处理进行分析,确定最少分组数,作为K-Means聚类算法的初试K值,由此作为聚类算法的起点,对交换机数量、空间布局以及网络结点分组进行设计。通过对QSKM算法和传统K-Means算法的计算复杂度进行分析比较以及QSKM算法在北京航天飞行控制中心显示网络设计中的应用研究表明,QSKM算法是有效的,降低了传统K-Means算法的计算复杂度。在我们设计的QSKM算法中,排队论可以为K-Means聚类运算中K的初始选取提供指导,聚类算法利用网络特征对基于排队论的网络设计方法进行完善,从而得到最佳的网络分布方案。

Queuing Theory Supervising K-Means Clustering Algorithm and ITS Application in Optimized Design of TTC Network

This paper proposed an improved K-Means clustering algorithm(QSKM) for optimization design of network.We analyzed the call/process by queuing theory and calculated the least network group,which was the initial K of K-Means clustring,so as to decide the number and the layout of switchers.We analyzed and compared the computational complexity of QSKM and traditional K-Means,and we also applied the QSKM on optimization design of network in BACC,which showed that QSKM algorithm was effective and reduced the computational complexity of traditional K-Means clustring algorithm.In the novel QSKM,queuing theory provided the guidence for selecting the initial K in K-Means clustering algorithm and the clustering helped the queuing theory with additional attributes to obtain the best network distributed.