Internet AS幂律建模及其参数估计

Methods to model and estimate scaling exponents of power-law for Internet autonomous system

为了精确建模Internet自治系统层面上的拓扑结构,提出了基于最小节点度和最大节点度的拓扑幂律模型及其参数估计新算法。针对Internet自治系统层拓扑实际测量数据,利用新算法对拓扑幂律模型中的最小节点度、最大节点度以及标度参数进行计算。实验结果表明,由新算法估计的Internet自治系统层拓扑幂律模型的最小节点度为1,最大节点度随网络规模的增大而增大,标度参数的误差与使用最大然似估计法误差一样均非常小,约为2.25。

In order to accurately model Internet topology on autonomous system（AS） level,a power-law model is improved based on the smallest and the largest node-degree,and a new algorithm of parameters estimation for the power-law model is developed.The smallest and the largest node-degree,the power-law parameter are estimated by the use of a new algorithm for the actual measurement data form Internet autonomous system.The experimental results show that the smallest node-degree is 1,the largest node-degree increases by the network size increasing,and the scaling exponent of power-law is 2.25,the error of the new algorithm is very small as the maximum likelihood estimation.