基于异构模型的铁路出发旅客衔接时间价值

Value of Access Travel Time for Railway Departure Passengers Based on Variation Structure Models

旅客时间价值是开展区域综合交通一体化研究的重要参数.本文针对现有文献中铁路旅客衔接时间价值研究较少、异构模型计算结果的比较不够充分等问题,以成都铁路东客站出发旅客衔接方式选择SP数据为依托,建立了MNL、ML、LCL、Hybrid等异构模型,对异构模型所得旅客个体和样本总体的等候时间价值、车内时间价值进行了比较.结果表明:模型结构对时间价值求解有较大影响,ML中单侧三角分布模型求解结果较为合理;Hybrid模型在拟合效果上表现优良,但在个体等候时间价值求解结果中呈现出较为明显的偏态性,给总体统计描述带来一定的困难.由于目前尚无关于时间价值的明确比选方法,应用中可考虑取异构模型计算结果的均值.

Value of travel time(VOT)is one of the fundamental parameters in regional integration transportation research.However,little attentions have been paid on railway access VOTs in literatures,and the comparisons on the VOTs derived from variation structure models have not been studied adequately.This paper focuses on the above mentioned questions based on the SP access mode choice survey data dedicated to Chengdu East Railway Station.Variation structure models such as MNL,ML,LCL and Hybrid are established,and the individual and overall sampled results of waiting time values and in-vehicle time values that are derived from above mentioned variation structure models are compared.The results indicate that:model structure has great effects on the outputs of travel time values;the one-sided triangular distribution based ML model seems to perform better than other ML models;while Hybrid model performs best in the goodness of model fit,it shows extreme skewness in individual waiting time values,which brings difficulties in its statistical descriptions.As there has no scope for selection of VOT values,an intuitively plausible way is to use the average value derived from variation structure models.