基于组合优化理论的无线网络流量建模与预测.docVIP

基于组合优化理论的无线网络流量建模与预测 　　摘 要： 无线网络流量受到上网成本、上网行为等因素的综合作用，具有随机性和周期性变化的特点，针对单一模型不能全面描述该变化特点的难题，提出基于组合优化理论的无线网络流量预测模型。首先采用自回归积分滑动平均模型进行建模，找出无线网络流量的周期性变化规律，然后采用相关向量机进行建模，找出无线网络流量的随机性变化特点，最后将它们的预测结果组合在一起进行单步和多步的无线网络流量预测实验。实验结果表明，该模型可以同时对随机性和周期性变化特点进行描述，预测精度高于单一自回归积分滑动平均模型或者相关向量机。 　　关键词： 无线网络； 自回归积分滑动平均模型； 建模与预测； 组合优化理论 　　中图分类号： TN92?34； TP391 文献标识码： A 文章编号： 1004?373X（2016）23?0043?04 　　Modeling and forecast of wireless network traffic 　　based on combinatorial optimization theory 　　CHEN Huafeng1， 2， LIU Jianing3 　　（1. School of Information Science and Technology， Hainan Normal University， Haikou 571158， China； 　　2. College of Qionghai Distance Education， Hainan Open University， Qionghai 571400， China； 　　3. Information Network and Data Center， Hainan Normal University， Haikou 571100， China） 　　Abstract： Since the wireless network traffic is synthetically affected by the factors of online cost and online behavior， it has the characteristics of randomness and periodic variation. To solve the difficulty that the single model can′t describe the change characteristic comprehensively， a wireless network traffic prediction model based on combinatorial optimization theory is put forward. The autoregressive integral moving average model is used to build the proposed model to find out the periodic variation rule of the wireless network traffic， the relevance vector machine is used to establish the model to find out the random variation characteristics of the wireless network traffic， and then the two prediction results are combined to realize the single step and multi?step wireless network traffic prediction experiments. The results show that the proposed model can describe the characteristics of randomness and periodic variation， and its prediction accuracy is higher than that of the single autoregressive integral moving average model or correlation vector machine. 　　Keywords： wireless network； autoregressive integral moving average model； modeling and prediction； combina