交通流时间序列的多重分形分析

Analysis of multifractal characteristic of traffic-flow time series

为研究交通流时间序列的分形特征,计算实测交通流量序列的多重分形谱,结果表明交通流序列的分形谱线为端点左高右低,且顶点左偏的曲线,说明序列的分形结构是不规则的,交通流量处于最大值的次数大于处于最小值的次数。比较不同统计间隔的交通流序列的多重分形谱,结果表明随着标度增大,分形谱宽度增加,但离差始终大于零,说明较大统计间隔的交通流序列的奇异性较强。

In order to analyze fractal characteristic of traffic-flow time series, multifractat spectrum of real traffic-volume time series is calculated.The result shows that left endpoint is higher than right endpoint,and vertex is in left side.It can be explained number of highest volume condition is larger than that of smallest.Compared with multifractal spectrum of traffic time series which have different statistic gaps, the gap is wider, and the width of spectrum is wider, but deviation remains larger than zero, so the traffic-volume time series which has large statistic scale has more multifractal characteristic and irregularity.