基于广义S变换和压缩感知的端到端流量估计算法（英文）  

End-to-End Traffic Estimation Algorithm Based on Generalized S-transform and Compressed Sensing

作　　者：杨济豪[1] 张自友[1] 姚春萍[2] 

机构地区：[1]乐山师范学院物理与电子工程学院,四川乐山614000 [2]东北大学信息科学与工程学院,辽宁沈阳110819

出　　处：《南开大学学报（自然科学版）》2016年第3期44-52,共9页Acta Scientiarum Naturalium Universitatis Nankaiensis

摘　　要：研究了高速网络中的流量矩阵估计问题.针对该问题在时域中的高病态特性,提出一种新的算法来解决该问题,该算法利用广义S变换和压缩感知来建模流量矩阵.通过广义S变换,则在时频域中对流量矩阵进行估计;在时频域中流量矩阵被分解为稳态分量和波动分量两部分,稳态分量通过抽样测量的样本数据进行建模分析,而波动分量则利用压缩感知理论进行估计.最后,利用真实网络流量数据验证所提出的算法,仿真结果表明所提出的算法是有效和可行的.The problem of traffic matrix estimation in high-speed network is studied. Because of the highly ill-posed nature of the problem in time domain, an new algorithm is proposed to overcome it, by combing general S-transform and compressive sensing to model the traffic matrix. By general S-transform,then the traffic matrix is estimated in time-frequency domain. The traffic matrix is divided into two partsthe stationary part and fluctuate part, with the stationary part being modeled by some measured sample data and the fluctuant part being estimated using CS theory. Finally, the true network flow data is used to verify the proposed algorithm and the simulation results demonstrate that the proposed algorithm is valid and effective.

关 键 词：流量矩阵 时频分析 广义S变换 压缩感知 流量估计 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]