Transfer Penalty Measurement of Intercity Rail Transit Hub Based on Nest Logit model
-
摘要: 在城市公共交通导向发展（TransiT-orienTed developmenT，TOD）模式下，衡量出行者在轨道交通枢纽的换乘便捷性是优化换乘组织设计的重要实践基础。然而，国内针对城际换乘阻抗的定量测算及不同群组间换乘阻抗差异的实证研究仍较为缺乏。因此，本研究构建巢式LogiT模型，分2层结构（支线出行与干线出行）刻画出行者的跨城交通方式选择行为。其中，干线出行方式包括私家车、高铁和长途汽车，支线出行方式涵盖步行/自行车、私家车/出租车/网约车，以及公共汽车/城市轨道交通。通过行为调查与意向调查采集居民城际出行方式选择行为数据进行模型参数标定。依据模型结果分析影响交通方式选择的关键因素，建立换乘阻抗的定量测算方法，并对比换乘阻抗在不同等级群组间的差异。结果表明：①城际交通方式选择会受到社会经济属性和出行方式特征的影响。在社会经济属性变量中，最显著影响因素包括：学历（t=3.492）、职业（t=3.422），以及是否拥有私家车（t=－5.722）；在出行特征变量中，最显著影响因素包括乘车时间（t=－4.745）和乘车费用（t=－5.935）；②经测算，出行者在出行中的车外时间价值、车内时间价值和延误时间价值分别为56.6元/h、55.0元/h以及58.0元/h；基于等效费用、等效车内时间和等效车外时间的换乘阻抗结果依次为25.4元/次、26.9 min/次，27.6 min/次。③不同学历、职业和收入群组间的换乘阻抗具有明显异质性。其中，硕士及以上群组的换乘阻抗约为本科群组的1.3倍，公职人员群组的换乘阻抗约是学生群组的3倍，高收入群组（＞10 000元）约为低收入群组（＞2 000~5 000元）的3.7倍。研究结果可为评估城际轨道交通枢纽的换乘效率和换成组织优化设计提供理论支持和实践参考。Abstract: In the context of transit-oriented development (TOD), assessing the transfer convenience of travelers at rail transit hubs serves as a fundamental basis for optimizing transfer organization and design. However, existing domestic research lacks quantitative evaluations of intercity transfer penalty and empirical studies on the differences in transfer penalty among different traveler groups. To address this gap, this study constructs a nested Logit (NL) model with a two-tier structure (feeder travel and trunk-line travel) to characterize travelers' intercity travel mode choices. The trunk-line travel modes considered include private cars, high-speed rail, and long-distance buses, while the feeder travel modes encompass walking/bicycling, private cars/taxis/ride-hailing services, and public buses/urban rail transit. To calibrate the model parameters, revealed preference and stated preference surveys were conducted to collect intercity travel mode choice data from residents. Based on the model results, key factors influencing travel mode selection were identified, a quantitative measurement method for transfer penalty was developed, and differences in transfer impedance across different demographic groups were analyzed. The main findings of this study are as follows: ①Intercity travel mode choices are significantly influenced by socioeconomic attributes and travel mode characteristics. Among the socioeconomic attributes, the most significant factors include education level (t= 3.492), occupation (t=3.422), and private car ownership (t=－5.722). Regarding travel characteristics, the most influential factors include travel time (t=－4.745) and travel cost (t=－5.935). ②The estimated value of time for different travel stages is as follows: out-of-vehicle time 56.6 CNY/h, in-vehicle time 55.0 CNY/h, and delay time 58.0 CNY/h. The transfer penalty values, based on equivalent cost, equivalent in-vehicle time, and equivalent out-of-vehicle time, are 25.4 CNY per transfer, 26.9 minutes per transfer, and 27.6 minutes per transfer, respectively. ③There is significant heterogeneity in transfer penalty among different education, occupation, and income groups. Specifically, the transfer penalty for the postgraduate group is approximately 1.3 times that of the undergraduate group; the public sector employees exhibit transfer penalty approximately three times higher than that of students; and the high-income group (>10 000 CNY/month) has a transfer penalty 3.7 times higher than the low-income group (2 000—5 000 CNY/month). These findings provide theoretical insights and practical guidance for evaluating transfer efficiency at intercity rail transit hubs and optimizing transfer organization and planning.
-
Key words:
- railway transportation /
- transit hub /
- transfer penalty measurement /
- Nest Logit model /
- TOD /
- group differences
-
表 1 模型变量定义表
Table 1. Model variable definition table
模型 影响因素 变量代码 支线模型 出行费用 C 等待时间 Tw 乘车时间 Tr 干线模型 性别 男性（取1） gen 年龄/岁 ＞45（取1） age 学历 本科（取1） edu1 硕士及以上（取1) edu2 职业 学生（取1） job1 企业职员（取1） job2 公务员（取1） job3 月收入/元 >2 000~5 000(取1) inc1 >5 000~10 000(取1) inc2 >10 000(取1) inc3 私家车 拥有私家车（取1） car 同行人数 有人同行（取1） par 行李件数 携带行李（取1 lug 车外时间 To 延误时间 Td 乘车时间 Tr 出行费用 C 换乘次数 N 
下载: 导出CSV
表 2 特征变量及水平值设置
Table 2. Characteristic variable and level value
特征变量 水平值 到达车站时间/min 15, 25, 35 Null 5, 15, 25 等待时间/min 15, 25, 35 Null 10, 15, 20 乘车时间/min 120, 150, 180 180, 210, 240 210, 240, 270 出行费用/元 150, 180, 210 300, 350, 400 70, 100, 130 换乘次数/次 1, 2, 3 0 1, 2
下载: 导出CSV
表 3 样本基本特征描述性统计
Table 3. Descriptive statistics of basic characteristics of samples
变量 分类 频数 百分比/% 性别 男 261 43.0 女 346 57.0 年龄/岁 $\leqslant 18$ 25 4.1 >18~30 277 45.6 >30~45 216 35.6 >45~60 80 13.2 >60 9 1.5 教育程度 本科以下 131 21.6 本科 353 58.1 硕士及以上 123 20.3 职业 学生 154 25.4 企业职员 294 48.4 公务员、事业单位 87 14.3 其他 72 11.9 月收入/元 $\leqslant 2~000$ 140 23.1 >2 000~5 000 164 27.0 >5 000~10 000 217 35.7 >10 000 86 14.2 常住地可使用私家车数量 无 210 34.6 1辆 352 58.0 2辆及以上 45 7.4
下载: 导出CSV
表 4 支线选择模型参数估计表
Table 4. Branch line parameter estimation
变量 系数值 t值 显著性 等待时间Tw -0.048 -1.333 乘车时间Tr -0.056 -2.280 * 出行费用C -0.128 -2.323 * 常数 0.318 0.445 Log-Likelihood: -332.74 拟合优度比ρ2=0.168 注：显著性列中，*代表水平为0.05；空格代表不显著。t值代表：对每1个自变量的逐个检验，如果t绝对值大于1.67就说明该变量前面的系数显著不为0，这个变量是有用的。
下载: 导出CSV
表 5 干线选择模型参数估计表
Table 5. Main line parameter estimation
变量类型 变量 系数值 t值 显著性 选择方案特征变量 车外时间To -0.008 -2.189 * 延误时间Td -0.009 -2.231 * 乘车时间Tr -0.008 -4.745 *** 出行费用C -0.009 -5.935 *** 换乘次数N -0.224 -1.920 · logsum 0.692 4.434 *** 个人属性变量 常数 长途汽车 -1.524 -2.631 · 高铁 -0.463 -0.921 gen 长途汽车 0.057 0.302 高铁 0.055 0.321 age 长途汽车 -0.296 -0.959 高铁 0.326 1.383 edu1 长途汽车 0.887 3.492 *** 高铁 0.237 1.052 edu2 长途汽车 0.883 2.696 ** 高铁 0.330 1.109 job1 长途汽车 0.353 0.928 *** 高铁 1.147 3.422 job2 长途汽车 0.344 1.075 高铁 0.483 1.769 · job3 长途汽车 -0.696 -2.471 * 高铁 -0.292 -1.261 inc1 长途汽车 0.643 1.827 · 高铁 0.948 3.032 ** inc2 长途汽车 0.295 0.790 高铁 0.716 2.190 * inc3 长途汽车 0.106 0.248 高铁 0.810 2.198 * car 长途汽车 -1.750 -6.991 *** 高铁 -1.333 -5.722 *** par 长途汽车 0.541 2.825 ** 高铁 0.258 1.574 lug 长途汽车 -0.388 -1.927 · 高铁 -0.176 -1.004 Log-Likelihood: -1 379 拟合优度比ρ2 =0.189 注：显著性列中，· 代表水平为0.1；*代表水平为0.05；**代表水平为0.01；***代表水平为0.001；空格代表不显著。t值代表：对每1个自变量的逐个检验，如果t绝对值大于1.67就说明该变量前面的系数显著不为0，这个变量是有用的。
下载: 导出CSV
表 6 时间价值及换乘阻抗计算统计表
Table 6. Time value and Transfer penalty calculation statistical table
测算内容 参数名称 数值 特征变量系数 车外时间 -0.008 3 延误时间 -0.008 5 乘车时间 -0.008 1 出行费用 -0.008 8 换乘次数 -0.223 6 车外时间价值（/元/h） 56.6 延误时间价值（/元/h） 58.0 乘车时间价值（/元/h） 55.0 换乘阻抗 等效车内时间（/min/次） 27.6 等效车外时间（/min/次） 26.9 等效费用（/元/次） 25.4
下载: 导出CSV
表 7 不同群组换乘阻抗计算统计表
Table 7. Transfer penalty for different groups calculation statistical table
分组依据 群组名称 换乘阻抗等效费用（元/次） 教育程度 本科 29.4 硕士及以上 37.3 职业 学生 10.3 企业职员 28.7 公务员 31.2 月收入（元）) >2 000~5 000 11.2 >5 000~10 000 23.8 >10 000 41.4
下载: 导出CSV
-
[1] 向红艳, 任小聪, 陈坚. 城市群内城际短途出行方式选择行为建模[J]. 重庆交通大学学报(自然科学版), 2016, 35(3): 129-133.XIANG H Y, REN X C, CHEN J. Modeling on choice behavior of short-distance intercity travel mode[J]. Journal of Chongqing Jiaotong University(Natural Science), 2016, 35(3): 129-133. (in Chinese) [2] ISEKI H., SMART M J. How do people perceive service attributes at transit facilities?[J]. Journal of the Transportation Research Board, 2012, 2274: 164-174. [3] DE K B, KOUWENHOVEN M, HOFKER F. New insights in resistance to interchange[J]. Transportation Research Procedia, 2015(8): 72-79. [4] GONG X L, CURRIE G, LIU Z Y, GUO X C. A disaggregate study of urban rail transit feeder transfer penalties including weather effects[J]. Transportation, 2018, 45: 1319-1349. doi: 10.1007/s11116-017-9768-0 [5] CHOWDHURY S, CEDER A. Users' willingness to ride an integrated public transport service: a literature review[J]. Transport Policy, 2016, 48: 183-195. doi: 10.1016/j.tranpol.2016.03.007 [6] CASCAJO R, LOPEZ E, HERRERO F, MONZON A. User perception of transfers in multimodal urban trips: a qualitative study[J]. International Journal of Sustainable Transportation, 2018, 13(6): 393-406. [7] KIM I, KIM H C, SEO D J. Calibration of a transit route choice model using revealed population data of smartcard in a multimodal transit network[J]. Transportation, 2020, 47(5): 2179-2202. doi: 10.1007/s11116-019-10008-8 [8] LIU R, PENDYALA R, POLZIN S. Assessment of intermodal transfer penalties using stated preference data[J]. Transportation Research Record, 1997, 1607: 74-80. doi: 10.3141/1607-11 [9] LUCAS M F, HENRIK B, KAY W A. Modelling intermodal travel in Switzerland: a recursive Logit approach. [J]. Transportation Research Part A: Policy and Practice, 2019, 119: 200-213. doi: 10.1016/j.tra.2018.11.009 [10] GUO Z, WILSON N H M. Assessment of the transfer penalty for transit trips in downtown boston a gis-based disaggregate modeling approach studies and planning[J]. Transportation Research Record, 2004, 1872: 10-18. doi: 10.3141/1872-02 [11] DOUGLAS N J, JONES M. Estimating transfer penalties and standardised income values of time by stated preference survey[J]. Australasian Transport Research Forum, 2013, 10: 1-21. [12] 关宏志. 非集计模型: 交通行为分析的工具[M]. 北京: 人民交通出版社, 2004.GUAN H Z. Disaggregate model-A tool of traffic behavior analysis[M]. Beijing: China Communication Press, 2004. (in Chinese) [13] 马莹莹, 陆思园, 张晓明, 等. 考虑个体风险偏好差异的高速公路出行选择模型[J]. 吉林大学学报(工学版), 2021, 51 (5): 1673-1683.MA Y Y, LU S Y, ZHANG X M, et al. Highway travel choice model considering individual risk preference differences[J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(5): 1673-1683. (in Chinese) [14] HENSHER D A, ROSE J M, GREENE W H. Applied choice analysis[M]. Cambridge: Cambridge University Press, 2015. [15] JOONHO K, SUGIE L, MIREE B. Exploring factors associated with commute mode choice: an application of city-level general social survey data[J]. Transport Policy, 2019, 75: 36-46. [16] PEREZ J B, NOVALES M, ORRO A. Spatially correlated nested Logit model for spatial location choice[J]. Transportation Research Part B: Methodological, 2022, 161: 1-12. [17] CHANDRA R B. An endogenous segmentation mode choice model with an application to intercity travel[J]. Transportation Science, 1997, 31(1): 34-48. [18] 张新洁, 关宏志, 赵磊, 等. 有限理性视野下出行者出行方式选择分层Logit模型研究[J]. 交通运输系统工程与信息, 2018, 18(6): 110-116.ZHANG X J, GUAN H Z, ZHAO L, et al. Study on the hierarchical Logit model of travelers' travel mode choice under bounded rationality perspective[J]. Journal of Transportation Systems Engineering and Information Technology, 2018, 18(6): 110-116. (in Chinese) [19] 刘建荣, 郝小妮. 基于随机系数Logit模型的支线出行方式选择行为研究[J]. 交通运输系统工程与信息, 2019, 19(5): 108-113.LIU J R, HAO X N. Study on urban travel mode choice behavior based on random coefficient Logit model[J]. Journal of Transportation Systems Engineering and Information Technology, 2019, 19(5): 108-113. (in Chinese) [20] 彭志鹏, 潘恒彦, 王永岗. 基于贝叶斯网络的网约车交通事故致因机理分析[J]. 东北大学学报(自然科学版), 2023, 44 (1): 145-152.PENG Z P, PAN H Y, WANG Y G. Analysis of the causes of traffic accidents of online car-hailing based on Bayesian network[J]. Journal of Northeastern University(Natural Science Edition), 2023, 44(1): 145-152. (in Chinese) [21] QIAN L, ZHUO L T, JIANG P Z. Assessing equitable accessibility based on total travel cost and multiple options: case study of airport ground access in Hong Kong[J]. Journal of Urban Planning and Development, 2022, 148(1): 04021061. -
点击查看大图
图(1) / 表(7)
计量
- 文章访问数: 16
- HTML全文浏览量: 6
- PDF下载量: 0
- 被引次数: 0