基于Edgeworth级数的行程时间可靠性的计算与实证

Calculation and Empirical Analysis of Travel Time Reliability Based on Edgeworth Series

通过引入行程时间可靠性概念,研究因受交通事故、交通供需矛盾等因素的影响,而导致路网行程时间随机波动的现象。为了研究路段中行程时间的分布规律,基于连续3个月的历史行程时间数据,实证分析了几个典型时间窗的行程时间分布形态,对高峰时段的数据拟合正态分布,平峰时段的数据拟合对数正态分布。同时,对大多数不服从常用理论分布函数的路段行程时间,使用Edgeworth渐进级数来近似其概率密度函数和分布函数,进而构建出相应的路段行程时间可靠性的计算模型。最后,结合高峰时段实证数据计算路段行程时间可靠性,并将其与基于正态分布计算出来的行程时间可靠性进行比较,实证结果表明基于Edgeworth级数估计的行程时间可靠性明显优于正态分布,且该方法具有较好的移植性。

By introducing the concept of travel time reliability, how traffic accidents and contradiction between traffic supply and demand influence road network travel time and cause it to fluctuate randomly is studied. In order to investigate the distribution law of section travel time, based on the historical travel time data of 3 consecutive months, an empirical analysis of the travel time distribution patterns of several typical time windows is conducted, a normal distribution and a logarithmic normal distribution are fit to the data of peak times and off-peak times respectively. Meanwhile, for the most road travel times that do not comply with common theoretical distribution functions are approximated by the Edgeworth asymptotic series for their probability density functions and distribution functions, thereby a calculation model of the travel time reliability of corresponding roads is constructed. Ultimately, the road travel time reliability is calculated combining with the empirical data of peak hours, and it is compared with the travel time reliability which computed by estimated by portability. normal distribution. The empirical analysis result demonstrates that the travel time the Edgeworth series is better than that by normal distribution, and this method reliability has good portability.