交通运输系统工程与信息

交通运输系统工程与信息 ›› 2014, Vol. 14 ›› Issue (5): 37-42.

• 智能交通系统与信息技术 • 上一篇    下一篇

基于浮动车数据的随机路径选择模型参数估计

李军*1，赵长相1，唐晓宇1，谢良惠2   

  1. 1.中山大学工学院，广州510006 2. 济南市城市交通研究中心，济南250101
  • 收稿日期:2014-01-22 修回日期:2014-05-06 出版日期:2014-10-25 发布日期:2014-12-17
  • 作者简介:李军(1968-)，男，湖北江陵人，副教授，博士.
  • 基金资助:

    国家自然科学基金资助项目(51178475).

Parameter Estimation of Stochastic Route Choice Model with Floating Car Data

LI Jun1，ZHAO Chang-xiang1,TANG Xiao-yu1, XIE Liang-hui2   

  1. 1.School of Engineering, Sun Yat-sen University, Guangzhou 510006, China； 2.Jinan Urban Transport Research Center，Jinan 250101, China
  • Received:2014-01-22 Revised:2014-05-06 Online:2014-10-25 Published:2014-12-17

PDF

被引次数

1

摘要：

提出利用浮动车出行数据对城市路网多OD对logit 随机路径选择模型参数估计方法.首先采用相对阻抗，实现logit 模型参数无量纲化和单一化，将离散多项logit 模型转化为相对阻抗的连续概率模型；然后对多OD对的浮动车数据进行分区和权重处理，计算得到概率分布；最后在推导连续型概率密度和累积概率函数基础上，利用最小二乘法进行参数估计.以广州市浮动车数据对方法进行验证，结果拟合优度高，所得参数可有效反映实际路径选择行为.方法简单实用，特别适用于利用海量浮动车数据进行路径选择行为分析.

关键词: 交通工程, 参数估计;浮动车数据;路径选择模型;相对阻抗

Abstract:

This study proposes an algorithm to estimate the parameter of Logit route choice model in urban road network with multinomial OD pairs utilizing taxi-floating car data. The relative route impedance is used and the Logit diversion parameter for whole network can be described by one unified dimensionless parameter. The discrete route choice probabilities are converted to continuous form with consideration of the weight of the numbers of trips, and a cumulative function relationship between route choice probability and relative impedance is re-derived from the original Logit route choice model. The least squares method is utilized to estimate of the unified parameter. The numerical example from taxi GPS data of Guangzhou shows that the goodness fit is excellent and the proposed algorithm is effective and convenient to model the route choice behavior with available and massive floating car data.

Key words: traffic engineering, parameter estimation, floating car data, route choice model, relative impedance

中图分类号: