基于TransCAD的反推OD矩阵敏感性分析

发布时间：2018-04-10 21:47

本文选题：反推OD矩阵 + TransCAD软件　；参考：《苏州科技大学》2017年硕士论文

【摘要】：随着交通信息采集技术的发展以及监控系统的普及,路网上路段流量的获取变得越来越容易。因此,通过路段观测流量来反推OD矩阵的方法日益受到重视。本文首先介绍了OD矩阵反推的相关理论和方法,并结合苏州市部分特定的城市路网建立了路网模型。介绍归纳了敏感性分析法,对影响OD矩阵反推结果的因素进行筛选分析,选出主要的影响因素:交通分配方法(用户平衡法和随机用户平衡法);BPR路阻函数中的α、β参数。结合反推目的,选取均方根误差、相对误差以及模态置信准则(MAC)作为敏感性分析评价指标,并对OD矩阵反推结果敏感性分析进行方案流程设计。通过标定α和β的基础值得出基准OD矩阵,利用单因素敏感性分析法分析影响因素,逐个变动因素得到不同的OD矩阵和路段分配流量,利用评价指标将反推OD矩阵与基准OD矩阵比较,路段分配流量与路段观测流量比较。结果表明,随机用户平衡法的反推精度大于用户平衡法,建议在反推OD矩阵时使用随机用户平衡法;在随机用户平衡方法下,路阻函数中α、β参数的取值会影响反推OD矩阵的精度,并提出了α和β的建议取值范围。研究成果可为更好地利用Trans CAD软件进行OD矩阵的反推、提高反推精度提供参考。
[Abstract]:With the development of traffic information collection technology and the popularization of monitoring system, it becomes more and more easy to obtain traffic flow on road network.Therefore, more and more attention has been paid to the method of backward extrapolating OD matrix through traffic flow observation.This paper first introduces the related theories and methods of OD matrix backstepping, and establishes the road network model combined with some specific urban road networks in Suzhou.The sensitivity analysis method is introduced in this paper. The factors influencing the inverse result of OD matrix are screened and analyzed, and the main influencing factors are selected: the traffic assignment method (user balance method and random user balance method) and the 伪, 尾 parameters in the BPR road resistance function.The RMS error, relative error and modal confidence criterion (MAC) are selected as the sensitivity analysis indexes, and the sensitivity analysis of OD matrix backstepping results is designed.By calibrating the basis of 伪 and 尾, the reference OD matrix is worth getting. The influence factors are analyzed by single factor sensitivity analysis method, and different OD matrices and road flow distribution are obtained one by one.The inverse OD matrix is compared with the reference OD matrix, and the section distribution flow is compared with the observed flow by using the evaluation index.The results show that the backstepping accuracy of the stochastic user balance method is greater than that of the user balance method, and it is suggested that the random user balance method should be used when the OD matrix is pushed back, and the random user balance method should be used in the random user balance method.The value of 伪, 尾 parameters in the road resistance function will affect the accuracy of the backstepping OD matrix, and the suggested range of 伪 and 尾 values is proposed.The research results can provide a reference for the better use of Trans CAD software to push back the OD matrix and improve the precision of the backstepping.
【学位授予单位】：苏州科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U491.12

【参考文献】