快速路交通流时间序列的多重分形分析  

Analysis of mul-tifractal properties of traffic flow time series on expressway

作　　者：董军 张勇 

机构地区：[1]中电投物流有限责任公司,北京100052 [2]武汉轻工大学数学与计算机学院,武汉430024

出　　处：《计算机工程与应用》2016年第10期227-230,共4页Computer Engineering and Applications

基　　金：国家自然科学基金(No.61201452;No.61179032)

摘　　要：采用多重分形消除趋势波动分析法研究城市快速路交通流的多重分形特征。首先,计算和比较了实测的流量和速度序列的广义hurst指数,结果表明:流量序列和速度序列均呈现多重分形,且流量序列的波动程度强于速度序列。然后,对两种序列的重排序列的计算结果显示,重排序列的广义hurst指数远小于原序列且不为0.5,说明交通流的多重分形特征主要由序列的长程相关性决定,同时与概率分布有关。In order to analyze the complex characteristic of traffic flow on expressway, multifractal detrended fluctuation analysis method have adopted to calculate the general hurst index of traffic flow time series. The result show that the capacity time series and speed time series are multifractal, capacity time series has more fluctuation than speed time series. The general hurst indexes of rearranged time series which disrupt the order of original series have been calculated, the rearranged general hurst indexed are not 0.5 and far smaller than the original series, then, the multifractal characteristic most decide by the long distance relevance, and has certain relation with the probability density.

关 键 词：多重分形 多重分形消除趋势波动分析 交通流时间序列 广义HURST指数 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] U491.112[自动化与计算机技术—计算机科学与技术]