【作者】 陈坚；

【导师】 晏启鹏；

【作者基本信息】 西南交通大学 ， 交通运输规划与管理， 2012， 博士

【摘要】 城市公共交通是具有“公益性”的重要基础设施,与群众生产生活、经济发展、社会稳定息息相关,直接影响城市整体功能的发挥、群众生活质量的提高和城市可持续发展。居民出行行为是公交定价的关键因素之一,而公交票价水平在服务质量一定的情况下,对出行行为结果将产生显著作用,也将决定公共交通的分担率,从而影响公交运营企业的收益和社会福利的共享。目前,我国公交定价依旧是静态的以政策调控和企业成本、利润为基础的定价模式,没有将考虑潜变量的城市居民出行行为纳入动态定价模型的建构中,公交时间差别定价也缺少研究,难以适应公交优先发展的要求。基于此,本文以大城市常规道路公交为研究对象,从公交出行行为分析与定价模型两方面入手,在完善及改进现有模型的基础上,研究二者的相互作用关系,从而构建出行行为与公交定价的整合模型。本论文的主要研究内容包含以下几个方面：1、在梳理城市交通系统范畴的基础上,扩展城市公共交通系统的定义,从系统参与者(政府管理部门、公交企业、乘客)三方来划分系统的构成。分析公交定价对城市交通系统结构的影响,根据传统城市道路交通网络流量分配模型,提出“方式—路径”图的概念,将其运用于描述出行全过程,构建基于网络均衡的城市交通系统客流分配模型,并设计了一个算例验证模型的有效性。2、运用计划行为理论(TPB)构建涉及出行者态度、主观规范及知觉行为控制等潜变量的出行行为分析模型,设计各潜变量的测量题项,揭示了潜变量在出行选择过程中的作用机理。实例分析结果表明：出行者知觉行为控制对出行行为意向有显著影响且为正相关,其次是出行者态度对出行行为意向的影响,而主观规范对出行行为意向有影响但影响最弱。3、从消费者行为理论中的EMB模式角度分析了出行方式选择行为的决策过程,提出了方式选择行为影响因素范畴(influence factor categories, IFC)的概念,将影响因素分为潜变量和显变量,并证明了因素间“黑盒子”的存在及求解思路。构建了出行方式选择行为结构方程模型,并对模型验证所需的调查问卷设计、信效度检验、求解步骤等进行了论述,并将模型运用于成都市公交出行意愿调查实例分析中。研究结果表明：知觉价值对行为意向的作用效果为0.84,而服务品质对知觉价值的作用效果为0.65,高于票价合理性的0.26,说明出行者对公交服务水平的看重高于价格,服务环境变量在服务品质中最为重要,其路径系数为0.68。最后根据出行者人格特质(personality traits)中的自我意识程度,及职业、收入作为分群标准,对不同群体的公交方式选择行为特性进行了定量对比分析。4、针对传统Logit方式选择模型效用函数仅考虑可观察的方案属性和个人社会经济属性,而没有加入影响出行者选择的心理因素这一不足,结合结构方程模型(SEM)与Logit模型构建了出行方式选择行为SEM—Logit整合模型,以改进已有研究中未能在方式选择模型中考虑潜变量的问题。并设计了整合模型的两阶段算法,从理论和应用两方面对模型的解进行证明,加入了服务品质满意度之后的整合模型,解释能力较传统Logit模型得到提升。5、从出行行为与公交定价二者相互作用机理出发,构建了多方式弹性需求下公交定价双层规划模型,有效的描述了公交网络系统票价水平变化对乘客出行总量、方式划分及路径选择等产生的影响。上层模型分别以企业利润最大化、乘客出行成本最小化和社会福利最大化为目标函数,下层模型为多方式弹性需求随机用户配流模型。算例结果表明：运用双层规划模型所确定的动态公交定价较传统静态定价能使政府、企业及出行者三方都获得更高收益,且上层模型以社会福利最大化为目标函数能代表社会群体中多数人利益,优化效果最为理想。6、根据城市交通出行时间不均衡性的特点,提出了时间差别定价的概念。并对公交时间差别定价的可行性进行了一定分析,构建了上层为社会福利最优模型,下层通过“非必要性”出行系数描述多方式分时段城市交通网络随机弹性需求的公交时间差别定价模型。模型整体采用改进遗传算法进行求解,下层多方式分时段弹性需求下随机用户平衡模型采用对角化算法与MSA算法组合求解。算例结果表明：时间差别定价方案比高峰/平峰同一费率定价方案社会福利目标函数高37.1,企业收益高36.5,公交客流高峰时段与平峰时段的比值由1.75：1降到1.16：1。并且实行高峰增加票价同时平峰降低票价的方案比仅对高峰时段增加票价或仅平峰时段降低票价效果最为明显,而乘客对平峰时段降低票价的弹性高于对高峰时段增加票价的弹性。综上所述,论文对城市公共交通基础理论、基于TPB理论的出行行为、出行方式选择行为结构方程模型、出行方式选择SEM-Logit整合模型、出行需求与公交票价双层规划模型及公交时间差别定价法等进行了全面深入的探索和研究,可为城市公共交通需求分析与票价制定等提供理论及方法支撑。更多还原

【Abstract】 Urban public transportation, a critical infrastructure of public service, which is closely related to citizen production and daily life, economic development and social stability, affects directly the overall function of the city, citizen life and urban sustainable development. Travel behavior is the key factor in urban public transit pricing, and transit pricing will play a significant role to the result of travel behavior, and it will determine public transport share rate and then will affect the survival of operating enterprises and the share of social welfare when service quality is unchanged. At present, bus fare is still a static pricing model based by policy regulation, corporate costs and profit, without considering the latent variable arising from travel mode choice behavior of urban residents into the construction of the dynamic pricing model. Moreover, temporal differentiatioan fare is hardly researched, delaying the requirement of bus priority development.Based on this, transit pricing model based on travel mode choice behavior was constructed by researching conventional bus transportation action on bus travel behavior analysis and pricing model establishment to improve the existing model and study the interaction between them. The main contents of this paper include:1. The definition of urban public transit system was extended, which included system participants of government departments, public transport enterprises and passengers, based on studying a category of urban transport system. The impact of public transit pricing on the structure of urban transport systems was analyzed, and then a concept of "model-route" graph was proposed to describe the whole process of travel according to traditional traffic network assignment model. At last, trip assignment model for urban traffic system based on network equilibrium was built and a numerical example was designed to verify the validity of the model.2. Travel behavior analysis model involved latent variables such as travelers’attitude, subjective norms and perceived behavioral control was constructed using the theory of planned behavior, and measurement items for each latent variable was design to reveal mechanism of the latent variable in the travel selection process. Example results show that travelers’perceived behavioral control have a significant impact on travel behavior intentions and have a positive correlation, followed by travelers’attitude on the impact of behavioral intentions, then the subjective norms have the weakest effect on intention of travel behavior.3. The concept of influence factor categories of travel mode choice behavior by analysis of travel mode choice behavior of decision-making process on the theory of consumer behavior patterns in the angle of EMB (Engel Miniard Blackwell model), which made the factors divided into the latent variables and variables and proved the existence of the "black box" idea between elements and given solution. And a structural equation modeling of public transportation mode selection behavior was built and to be used in the analysis of Chengdu bus travel survey examples, and then questionnaire designing, test on reliability and validity and solving steps required for model validation were discussed. The results show that perceived value on the effect of behavioral intention is quantified to be0.84, service quality on the effect of perceived value is quantified to be0.65, greater than0.26of the price rationality, which illustrate that travelers think level of service are valued higher than the price, service environment variables is the most important one in service qualities whose path coefficient is0.68. Finally, taking self-awareness level of travelers personality traits, occupation, income as a grouping criteria, a quantitative comparative was analyzed for different groups of public transport mode selection behavior characteristics.4. For traditional utility functions in Logit mode choice model considered only the observable program properties and individual socio-economic attributes, but did not consider psychological factors which affect travel choices, SEM-Logit integration model of bus mode choice behavior was built combined with structural equation modeling (SEM) and Logit model to solve the problem that current studies didn’t consider latent variable in travel mode selection model. And two-phase algorithm of integrated model was designed, whose solution was proved by both theory and application that the ability to explain was promoted more when considering satisfaction of service quality in integration model than traditional Logit model.5. Bi-level programming model of public transit pricing based on Multi-mode network with elastic demand was constructed from interaction mechanism between travel behavior and public transport pricing, which effectively described that public transport network fare changes have an impact on the total travel passengers, mode split and route choice. The objective function of the upper model is to maximize their profits, to minimize the cost of passenger and to maximize social welfare, the lower was a random user assignment model based on Multi-mode network with elastic demand. Numerical results showed that dynamic bus pricing using bi-level programming model can obtain higher yields for government, enterprises and travelers than the traditional static pricing, and the upper model took social welfare maximization as the objective function can represent the interests of the majority social groups and made the effect of optimize most satisfactory.6. The concept of temporal differentiation fare was proposed according to characteristics of unbalanced nature urban transport travel time and the feasibility was analyzed. Then the temporal differentiation fare mode was constructed, the upper model is a model with optimal social welfare and the lower model described Multi-mode and Different Times network with elastic demand by unnecessary travel coefficient. The overall model was solved using improved genetic algorithm and the lower model was solved by diagonalization algorithm combined with the MSA algorithm. Numerical results show that the social welfare objective function of temporal differentiation fare program was37higher, and corporate earnings was36.5higher than the ones of the same flat rate pricing program in the peak level, the ratio was reduced from1.75:1in bus passenger flow peak periods to1.16:1in the flat peak periods. Another conclusion was that the program of increasing fare in the peak level while reducing fare in the flat peak had a more obvious effect than a program of only increasing fare in the peak level or a program of only reducing fare in the flat peak, and that passengers have a greater elasticity when reducing fare in the flat peak than increasing fare in the peak level.In summary, the paper fully involved in the exploration and research on basic theory of urban public transport, TPB-based theory of travel behavior, structural equation modeling of travel mode choice behavior, SEM-Logit model integration of transportation mode selection, bi-level programming model of travel demand and transit pricing, which can provide the theory and methods for analysis of urban public transport demand and public transit pricing.更多还原