计算机科学 ›› 2015, Vol. 42 ›› Issue (4): 68-71.doi: 10.11896/j.issn.1002-137X.2015.04.012

• 网络与通信 • 上一篇    下一篇

基于定量递归联合熵特征重构的缓冲区流量预测算法

陆兴华,陈平华   

  1. 广东工业大学华立学院 增城511325,广东工业大学计算机学院 广州510006
  • 出版日期:2018-11-14 发布日期:2018-11-14
  • 基金资助:
    本文受广东省教育部产学研结合项目(2012B091100003,2B091000058)资助

Traffic Prediction Algorithm in Buffer Based on Recurrence Quantification Union Entropy Feature Reconstruction

LU Xing-hua and CHEN Ping-hua   

  • Online:2018-11-14 Published:2018-11-14

摘要: 对网络基站缓冲区的短时网络流量的准确预测是缓解和控制拥堵的关键。基站缓冲区的短时网络流量时间序列具有非线性混沌特征,其自相关特性较弱,而传统方法采用线性时间序列分析方法没能有效挖掘流量序列的非线性特征信息,流量序列预测性能不好。提出了一种基于非线性时间序列分析的定量递归联合熵特征重构网络基站缓冲区的短时网络流量预测算法,该算法提取流量序列的定量递归联合熵特征,并对特征序列进行相空间重构;把网络流量信号模型进行高维映射,在高维相空间对短时网络流量序列进行定量递归分析;采用自相关特征奇异分解对流量序列进行聚合后的线性叠加,采用平均互信息算法和虚假最近邻点算法计算最佳时延参数和最小嵌入维;进行插值拟合形成时频分析特征分流控制,实现对网络流量的预测。仿真结果表明,该算法预测精度较高,稳定性较好,预测偏差较传统方法低,具有较好的应用价值。

关键词: 基站缓冲区,网络流量,预测,非线性特征

Abstract: The accurate prediction of short-time traffic network in base station buffer is the key to alleviate and control congestion.The short-time traffic network traffic has nonlinear chaotic characteristics,and the self correlation is weak.The traditional method uses linear time series method,and the nonlinear feature information is not used,so the prediction performance is not good.An improved traffic prediction algorithm based on nonlinear time series analysis and recurrence quantification union entropy feature reconstruction was proposed.The union entropy feature is extracted.The phase space reconstruction of characteristic sequences is obtained.The signal is mapped in the high dimensional phase space,and the recurrence quantitative analysis is taken for the traffic series.The autocorrelation characteristic singular decomposition is used for linear superposition after polymerization on the runoff series.The average mutual information method and false nearest neighbor algorithm are used for parameters optimization.Interpolation is taken for time frequency analysis and traffic flow control.The prediction of network traffic is completed.Simulation results show that this algorithm has high prediction accuracy and good stability,and prediction error is lower than traditional method,which shows good application value.

Key words: Base station buffers,Network traffic,Prediction,Nonlinear feature

[1] 宋杨,涂小敏,费敏锐.基于FARIMA模型的Internet时延预测[J].仪器仪表学报,2012,33(4):757-763
[2] 温祥西,孟相如,马志强,等.小时间尺度网络流量混沌性分析及趋势预测[J].电子学报,2012,40(8):1609-1616
[3] 朱凡,吴敏.基于定量递归分析的校园网流量特性分析[J].计算机应用与软件,2012,9(6):275-281
[4] 杨雷,李贵鹏,张萍.改进的Wolf一步预测的网络异常流量检测[J].科技通报,2014,30(2):47-49
[5] 张宾,杨家海,吴建平.Internet 流量模型分析与评述[J].软件学报,2011,2(1):115-131
[6] 许利军,杨棉绒.网络数据流量组播路由的多种群遗传算法[J].科技通报,2012,8(5):171-175
[7] Gross J,Janke W,Banchmann M.Massively parallelized replica-exchange simulations of polymers on GPUs [J].Comput.Phys.Comm.,2011,182:1638-1644
[8] 张凤荔,赵永亮,王丹,等.基于流量特征的网络流量预测研究[J].计算机科学,2014,1(4):86-89
[9] 黎峰,吴春明.基于能量管理的网络入侵防波动控制方法研究[J].计算机仿真,2013,0(12):45-48,335
[10] 张骏,田泽,梅魁志,等.基于节点预测的直接Cache一致性协议[J].计算机学报,2014,37(3):700-720
[11] 张萌,张沪寅,叶刚.延迟时间和嵌入维数联合优化的网络流量预测[J].计算机工程与应用,2014,0(4):103-109

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!