交通时间序列的多重分形特征分析

Multifractal analysis of traffic flow time series

通过分析2006年8月1日0时11分到8月10日14时11分的北京玉泉营高速路的交通流量时间序列,发现其具有多重分形特征。首先利用幂谱、统计矩的方法研究流量时间序列表明了该序列自相似性。进一步借助多重分形谱的宽度、极大值以及谱曲线不对称程度对交通流量时间序列进行研究,结果表明,交通流量时间序列具有较好的多重分形特性。这将为多重分形在交通时间序列分析的研究提供重要的理论基础。

This paper analyzes the time series of the traffic flow at Yuquanying freeway of Bejing from No. 1 to 10 August 2006 and author finds the multifractal characters. The self- similarity of the time series of the traffic flow are found by using the method of power spectrum and statistical moments. The width, max. and asymmetry parameter of multifractal spectrum are employed in studying the traffic flow time series , which shows the strong degree of multifractality in the time series. This will provide an important theoretic foundation for researching muhifractality in the theory of traffic time series.