申请同济大学工学博士学位论文


出行者对城市交通拥堵影响评价及行为反应研究

  (国家自然科学基金重点项目:信息环境下的城市道路交通规划理论与方法  编号:50738004)

 

培养单位:交通运输工程学院

一级学科:交通运输工程

二级学科:交通运输规划与管理

研 究 生:叶 亮

指导教师:杨东援 教授

二○一○年六月

 

A dissertation submitted to

Tongji University in conformity with the requirements for

the degree of Doctor of Philosophy

Research on Travel Demand Analysis of Urban Public Transportation

Based on Smart Card Data Information

  (Supported by the Key National Natural Science Foundation of China, Urban Transportation Planning Theory and Methods under the Information Environment  Grant No: 50738004)

 

School: School of Transportation Engineering

Discipline: Transportation Engineering

Major: Transport Planning and Management

Candidate: Ye Liang

Supervisor: Prof. Yang Dong-yuan

June, 2010

 

摘  要

交通拥堵问题严重影响着城市的运转效率和经济发展,制约着居民的生产和生 活出行。交通研究者和管理者致力于从交通规划、交通基础设施建设、提高交通系 统运力等硬对策和交通管理、需求引导、政策措施等软对策两个方面,寻求缓解交 通拥堵的对策方法。出行者作为交通系统的使用者,其思想、感受、意愿和行为是 评价交通系统服务水平、决定交通系统运行状态的关键因素。

在交通拥堵成为常态的情况下,如何合理设定城市交通系统服务水平目标已经 成为亟待解决的重要问题。随着城市交通拥堵问题的加剧,交通拥堵对策研究已经 成为国内交通研究的重要领域。然而,这些研究在思路、方法和技术上并无突破性 进展,对拥堵对策的研究多为定性分析、对策提出,或对策实施前后的建模分析, 缺少对拥堵影响及政策效果的系统性研究。从出行者角度评价交通拥堵的影响,研 究出行者在拥堵状态下的行为反应和对交通拥堵软对策的接受度,可为拥堵成为常 态情况下城市交通系统服务水平目标的确定提供依据,也可为交通拥堵对策研究提 供基础研究支撑。

论文以交通行为学分析为主要方法,以上海市居民交通拥堵状态下的行为和意 愿调查数据为研究基础,从理性-经济人(Economic man)向复杂人(Complex man) 转变的角度研究了交通拥堵对出行者的影响及出行者对拥堵的行为反应和拥堵对策 的接受度。

总结和分析了国内外关于交通拥堵影响和评价的研究。梳理了交通调查技术的 发展历程。整理和归纳了基于出行者态度和意愿的交通行为研究、主客观数据融合 的交通行为建模在国内外交通领域的研究现状。介绍了计划行为理论和应用。阐述 了因子分析法、聚类分析法和离散选择模型3种交通行为分析相关模型方法。

以上海市居民交通拥堵状态下行为与意愿调查为例,详细说明了如何采用多种 随机抽样相结合的方法,解决样本的可代表性和推广性问题;如何运用网络调查和 邮寄调查相结合的方式,协调调查时间、成本与样本信息完整度的矛盾;如何设计 调查问卷;如何通过细节设计和试调查改进调查的实施和问卷的设计,以保证数据 的准确性;如何进行数据的整理和清洗;如何基于社会心理学的态度和意愿条目设 计方法,运用李克特量表法获得量化的态度、偏好和意愿数据。

提出了交通行为分析中的复杂人假设,分析了用于提取出行者的态度和意愿特 征的因子分析法的原理及运用,得到了 14 个表征出行者满意度、个性、态度意愿及交通偏好的特征因子。分析了对出行者群体进行细分的聚类分析法的原理及运用, 得到了基于态度意愿细分的 2组出行者群体和基于交通偏好的 2组出行者群体。统 计了出行者主观特征细分群体的客观特征(性别、年龄、职业等) ,建立了主观群 体划分与客观特征的对应关系。

从出行者角度出发,提出了三类交通拥堵主观测度指标。建立了 Ordered logit 模型和多元回归模型,用以分析影响出行者对拥堵感知的因素变量。影响因素模型 结果表明,社会经济属性、态度和意愿因子、人群细分变量、出行基本属性和交通 拥堵现状和感受变量都对基于出行者角度的拥堵的测度有着显著的影响。针对交通 行为意愿数据采集的阶段性和小样本的问题,提出需建立拥堵主观指标与连续数据 环境下的客观指标对应关系。

基于大量的文献综述性研究,提出了交通拥堵状态下出行者的行为反应类型, 并进行了归类;基于计划行为理论提出了出行者面对拥堵的行为反应机理,并以此 为研究脉络,展开了拥堵状态下出行者的行为意向和决策的统计分析。对缓解交通 拥堵对策的实施对象和影响因素展开了研究,将拥堵对策根据研究需要划分成与拥 堵行为反应相关和与交通系统或经济调节相关两类。对两类对策分别建立了 Nested logit模型和Ordered logit模型,得到对策的实施对象特征和影响因素。模型 结果表明个人态度和意愿因子与社会经济属性变量一样,是决定出行者行为反应和 对对策支持程度的重要因素。出行基本属性和实际的交通拥堵现状对公共交通政策 和增加高速公路政策有一定的影响,而对拥挤收费、电子办公和电子商务以及错峰 出行等政策的影响不大。

论文从交通拥堵对出行者的影响评价和出行者对拥堵的行为反应角度,研究了 提高交通行为与意愿数据准确性的调查方法,为以后类似的调查设计提供了借鉴; 态度、偏好和意愿数据的采集方法,出行者潜在态度意愿特征因子的提取和出行者 的群体细分方法,完善了出行者差异化的研究技术;建立了出行者对交通拥堵影响 的评价指标,为从公众角度评价交通系统服务水平提供了依据;分析了出行者面对 拥堵的行为反应和机理,评估了相关交通拥堵对策的可接受度,为交通拥堵对策的 制定提供了决策支持和基础研究支撑。

关键词: 交通行为调查、拥堵评价、行为反应、态度和意愿特征、离散选择 模型

 

ABSTRACT

Traffic congestion as a big problem for urban development impacts residents’ daily trips. There are two ways to relieve traffic congestion: one is to improve transportation capacity by transportation planning, infrastructure constructing, et al. The other one is to reduce traffic demand or to change traffic demand mode by transportation demand management, traffic congestion policies, et al. Transportation users is the key point for urban transportation system, who impact transportation system condition and can evaluate it as the customer.

As it is hard to eliminate traffic congestion, we need to find another way to evaluate the service level of transportation system. It is a very popular area to study of relieving traffic congestion policies in China, however, until now, there is no study has made remarkable progress. Most of these studies focus on seeking sources of congestion and qualitative analysis of policies to relieve congestion. Quantitative indicators for congestion impacts are needed, as well as the study for estimating congestion policies. To study the traffic congestion impacts on transportation users and their reactions can provide some supports for setting the target of urban transportation system service level, also for choosing congestion policies.

The study focuses on the congestion impacts on transportation users and their reactions, also estimates some congestion policies by travel behavior study methods. A survey was taken to collect data on residents living in Shanghai of their travel attitude and behavior under the congestion in August 2009. When the study started, a hypothesis was made that those transportation users are complex man instead of economic man hypothesis which was used on traditional studies.

A literature review was made on traffic congestion indicators study, survey methods development, attitude analysis in travel behavior study, and objective and subjective data using in travel behavior study. It also contains an introduction of Theory of planned behavior (TPB) and three study models--- factor analysis, cluster analysis and discrete choice model which related with travel behavior analysis.

In order to represent the total population, a random sampling was used in the survey. Considering both sides: restriction (time and cost) and requirements (complete and rich information), a mixed survey structure (internet plus mail) was determined. The study also shows how to improve survey quality by questionnaire design, changing question order, considering question words. A pilot survey was implemented to estimate response rate, improve survey design and questionnaire design. Data cleaning and coding is another way to improve data quality, which was also contained in the study. Attitude statements design is a way to get attitude data through questionnaire survey. And the “Likert-scale” method was used in the survey to estimate the attitudes.

Factor analysis was used to get attitude factors based on the complex man hypothesis. 14 factors could be used to represent transportation users’ satisfaction, personal characteristics, general attitude and transportation related attitude. Cluster analysis is then conducted to identify market segments having different attitudinal profiles respectively by general attitudes and travel-related attitudes. These segments are different on a number of sociodemographics, such as gender, age, occupations et al.

Three kinds of traffic congestion indicators were built up based on transportation users’ feeling. The ordered logit model and the multiple linear regression model were set up respectively to analyze factors which impact transportation users’ feeling about congestion. The results show that sociodemographics variables, attitude variables, segments variables, trip characteristics variables and traffic condition variables are all significant in models. A relationship was built up to present these congestion indicators based on users’ feeling by indicators based on real-time traffic detection data (such as speed, delay, et al).

A series of users’ reactions by congestion were listed based on a review of previous studies. The mechanism of reactions under the congestion was studied by using the Theory of planned behavior (TPB). Two kinds of congestion policies were estimated by nested logit model and ordered logit model respectively. The model results show attitude variables are significant in the model as well as sociodemographics which impact transportation users’ acceptation of congestion policies. However, trip characteristics variables and traffic conditions variables just work for developing public transportation policy and paying more taxes to build more highways policy, but no sense to congestion price, telecommuting, and change departure time.

Generally, the study analyzed traffic congestion impacts on transportation users and their reactions. A survey was designed to collect travel attitude and behavior data, which is a good example about how to improve data quality for the similar surveys. Factor analysis and cluster analysis were used in the study to get attitude factors and segments based on attitudes, which are proved as a good method to study transportation users’ characteristics. Traffic congestion indicators were set up based on transportation users’ feeling, which can be used to evaluate transportation system service level. In order to support traffic congestion policies determination, transportation users’ reactions to congestion were studied and models were built to estimate the acceptation for congestion policies.

Key Words:  travel behavior survey, traffic congestion indicators, behavior reactions, attitude characteristics, discrete choice model