Abstract: It is very difficult to estimate the Traffic Matrix (TM) of the network, because it is a highly ill-posed problem. To solve the problem, a traffic matrix estimation method based on the Ant Colony Optimization (ACO) algorithm was proposed. Through appropriate modeling, the traffic matrix estimation problem was transformed into the optimization problem, and then the model was solved by ACO algorithm, which could effectively estimate the traffic matrix. Through the test results, compared with the existing methods, the accuracy of proposed algorithm is a bit weaker than entropy maximization and quadratic programming. But these two methods have high complexity, and they cannot be applied to large-scale network. Therefore, in the large-scale network, the proposed algorithm is better. It can improve the accuracy of traffic matrix estimation.

Key words: IP network, Origin-Destination (OD) flow, Traffic Matrix (TM) estimation, Ant Colony Optimization (ACO) algorithm

摘要： 针对IP网络流量矩阵(TM)估计的高度病态性，导致很难精确估计网络流量矩阵，因此提出了一种基于蚁群优化(ACO)算法的IP网络流量矩阵估计方法。通过适当的建模，将流量矩阵估计问题转化为最优化问题，再通过蚁群算法求解模型，有效解决了网络流量矩阵估计。通过测试结果分析，与现存的方法相比，所提算法的精度比最大熵和二次规划稍差，但这两种方法复杂度太高，不适用于大规模网络，因此，在网络规模较大的情况下，算法是较优的，可提高流量矩阵估计的精度。

关键词: IP网络, 源〖CD*2〗目的流, 流量矩阵估计, 蚁群优化算法