交通运输系统工程与信息

交通运输系统工程与信息 ›› 2023, Vol. 23 ›› Issue (4): 290-297.DOI: 10.16097/j.cnki.1009-6744.2023.04.029

• 工程应用与案例分析 • 上一篇    下一篇

建成环境对城市轨道交通起讫点客流的非线性影响及阈值效应

许奇1a，李雯茜1a，陈越1a，胡佳俊2，梁肖*1b   

  1. 1. 北京交通大学，a. 综合交通运输大数据应用技术交通运输行业重点实验室，b. 中国综合交通研究中心，北京 100044； 2. 八维通科技有限公司, 杭州 311199
  • 收稿日期:2023-01-17 修回日期:2023-03-03 接受日期:2023-03-06 出版日期:2023-08-25 发布日期:2023-08-22
  • 作者简介:许奇(1982- )，男，云南普洱人，副教授，博士
  • 基金资助:
    国家自然科学基金(71621001, 72171021)

Nonlinear and Threshold Effects of Built Environment on Origin-destination Flows of Urban Rail Transit

XU Qi1a, LI Wen-xi1a, CHEN Yue1a, HU Jia-jun2, LIANG Xiao*1b   

  1. 1a. Key Laboratory of Transport Industry of Big Data Application Technologies for Comprehensive Transport, 1b. Integrated Transport Research Center of China, Beijing Jiaotong University, Beijing 100044, China; 2. Bwton Technology Co. Ltd., Hangzhou 311199, China
  • Received:2023-01-17 Revised:2023-03-03 Accepted:2023-03-06 Online:2023-08-25 Published:2023-08-22
  • About author:liangx@bjtu.edu.cn
  • Supported by:
    National Natural Science Foundation of China (71621001, 72171021)

PDF

摘要： 城市轨道交通起讫点(OD)客流与建成环境的依赖关系研究有助于 TOD(Transit-oriented development)模式的实施。既有研究多关注建成环境对进出站客流的影响，而基于OD客流的研究未充分考虑建成环境要素的交互效应对OD客流的影响。采用多源位置大数据系统地刻画城市轨道交通的TOD建成环境，基于极限梯度提升决策树模型(XGBoost)研究城市轨道交通OD客流与TOD建成环境的非线性关系。针对北京地铁的案例研究表明：XGBoost能有效地处理建成环境对OD客流的非线性影响，其解释能力达到72.6%，估计结果更为可靠。TOD建成环境因子对OD客流的影响差异显著。密度和公共交通可达性等两类要素的重要度排序前二，其解释变量的平均重要度达到4.41%和3.71%，是全部变量平均值的1.29倍和1.08倍。解释变量重要度排序高的建成环境因子对OD客流的影响存在非线性特征，表现为显著的阈值效应。基于双变量部分依赖图的分析表明，城市轨道交通客流的流动依赖于起讫点建成环境的差异及其引发的交互效应。因此，发展城市轨道交通TOD时，不仅需从交通生成角度分析建成环境对进出站客流的影响，还需考虑客流的矢量性，从交通分布角度研究建成环境各要素的资源协同配置问题。

关键词: 城市交通, TOD建成环境, OD客流, 机器学习模型, 阈值效应

Abstract: Studies on the dependency relationship between the origin-destination (OD) passenger flow of urban rail transit and built environment indicators are useful to enhance Transit-oriented development(TOD). Existing studies have extensively examined the effects of the built environment on ingress/egress passenger flow. The studies that investigate the impact of the built environment on OD passenger flow seldom consider the effects of the interaction of built environment determinants on OD passenger flow. To this end, this study first uses multi-source location-based big data to describe the indicators of the TOD built environment of urban rail transit and then applies the extreme gradient boosting model (XGBoost) to investigate the nonlinear relationship between OD flows of the Beijing subway and TOD built environment. The case study of Beijing indicates that XGBoost can be able to identify the nonlinear relationship between OD flows and the TOD built environment indicators with a more reliable estimate result. Its interpretation ability reaches 72.6%. The difference in the effects of the built environment on OD flows is significant. The importance of two group indicators, namely density and public transport accessibility, ranked the top two. The average importance of indicators in these two groups is 4.41% and 3.71% , respectively, which make 1.29 and 1.08 times all variables' average values. The nonlinear effect of key indicators on OD flows is significant and shows a threshold effect. And the bivariate partial dependence diagram shows that the movement of urban rail transit passengers is determined by the difference in the built environment of OD pairs and the corresponding interaction between them. Hence, developing TOD not only needs to examine the influence of the built environment on ingress/egress flow from the perspective of traffic generation but also needs optimization and coordination of land use around urban rail transit stations from the perspective of traffic distribution.

Key words: urban traffic, TOD built environment, OD flow, machine learning model, threshold effect

中图分类号: