交通运输系统工程与信息 ›› 2013, Vol. 13 ›› Issue (2): 158-163.

• 系统工程理论与方法 • 上一篇    下一篇

Logit随机网络配流模型的改进Dial算法

杨泳,严余松*,户佐安,马毅   

  1. 西南交通大学 交通运输与物流学院, 成都 610031
  • 收稿日期:2012-10-30 修回日期:2013-01-25 出版日期:2013-04-25 发布日期:2013-04-27
  • 作者简介:杨泳(1982-),男,湖北大治人,博士生.
  • 基金资助:

    国家自然科学基金(61104175).

Improved Dial’s Algorithm for LogitBased Stochastic Traffic Assignment Model

YANG Yong, YAN Yu-song, HU Zuo-an, MA Yi   

  1. School of Transportation and Logistics, Southwest Jiaotong University, Chengdu 610031, China
  • Received:2012-10-30 Revised:2013-01-25 Online:2013-04-25 Published:2013-04-27

摘要:

研究Logit随机网络配流模型及实现模型求解的Dial算法,针对原模型及算法的缺陷,通过引入路段长度相关的容错系数指标重新定义有效路径的判定条件,在此基础上提出一种改进的Dial算法,并应用于Logit随机网络配流模型中.改进算法在不降低原算法精度下不仅保留了原算法的无需路径枚举、计算效率高等优越性,而且满足实际出行者偏好在较短路段上“迂回”选择潜在有效路段的特点.最后通过一个路网实例对2种算法的配流结果进行了对比.结果表明,改进的算法避免了原算法缺陷导致的结果异常,配流效果更加符合实际,其计算效果明显优于原算法.

关键词: 城市交通, 配流, 有效路段, Logit模型, Dial算法, 交通网络, 容错系数

Abstract:

This paper analyzes the multipath Logit assignment model and the classical Dial’s algorithm. To overcome the drawback in existing model and algorithm, an improved algorithm is presented. The efficient path condition is modified based on a quantitative correction coefficient on individual path. The improved algorithm retains the main advantages of original algorithm, in which the path enumeration is not required and a similar computing efficiency with the original algorithm is guaranteed. Meanwhile, the preference to potential valid road links with short distance among urban traffic travellers is satisfied. A numerical example based on a traffic road network is used to illustrate the application of proposed algorithm and two algorithms are then compared. Results show that the improved algorithm is able to eliminate the abnormal behavior of the original algorithm with more feasible and reasonable traffic flow assignment results. It is evident that the proposed algorithm is more effective than the original one.

Key words: urban traffic, traffic flow assignment, effective path, Logit model, Dial’s algorithm, traffic network, correction coefficient

中图分类号: