【作者】 张薇；

【导师】 何瑞春；

【作者基本信息】 兰州交通大学 ， 管理科学与工程， 2017， 博士

【摘要】 出租车合乘模式具有缓解城市交通拥堵、节约资源、减少污染的优势,许多国家已经实施了合乘制度,并取得了较好的效果,而我国各大城市也正在逐步推行出租车合乘制度以缓解各地的交通问题。目前我国正处于出租车合乘制度的起步阶段,合乘制度的实施过程中存在诸多亟待解决的问题。乘客与司机的行为对合乘制度的执行影响较大,现实中的很多乘客与司机拒绝合乘的现象十分普遍,这将严重阻碍合乘制度的实施。因此,对合乘制度下乘客与司机的行为研究显得十分必要。考虑到人类的非完全理性特点,本文在有限理性框架下,基于前景理论、演化博弈等理论方法,对出租车合乘行为问题展开了深入的研究,具体工作如下:(1)研究了出租车合乘效果,提出了一种合乘效果仿真分析方法。通过分析不同合乘情形下的乘客费用及司机收入,建立了出租车合乘模型,设计仿真算法,从乘客费用、司机收入及出租车市场的载客能力方面对合乘效果进行仿真,并分析了支付比例及交通拥堵率对合乘效果的影响,验证了合乘模式的有效性,得到了合乘效果的影响性规律。(2)研究了乘客的合乘行为,提出了一种乘客合乘决策的分析方法。以乘客的行程时间、费用及舒适度为指标,基于前景理论建立了乘客合乘决策模型,分析不同参照点下的乘客合乘行为以及乘客合乘决策的影响因素。提出的合乘决策分析方法兼顾了乘客的心理因素及交通状况的不确定性,体现了乘客的有限理性决策特点,是一种不确定环境下的乘客合乘决策的分析方法。(3)研究了司机的合乘行为,提出一种合乘模式下司机收入变化及其行为分析的方法。通过分析不同情形下司机的收入,考虑司机的公平心理,基于公平理论建立司机收入公平模型,研究了合乘模式下司机的平均收入及个体收入的变化,分析了司机收入变化的影响因素。所提出的司机收入公平模型考虑了司机的心理特点,分析了司机公平心理导致的行为变化,为有限理性下司机合乘行为分析提供了新的思路。(4)研究了乘客与司机的合乘策略,基于提出的乘客合乘决策方法深入研究了乘客在不同合乘情形下的策略选择,并将前景理论引入演化博弈中,改进了演化博弈模型,建立了乘客与司机的合乘策略博弈模型,利用复制动态机制进行演化分析,从而得到合乘稳定策略。该方法克服了演化博弈模型中完全理性的策略收益的问题,既考虑到合乘主体收益判断的有限理性,又兼顾了有限理性参与人策略的动态调整。(5)研究了合乘绕行问题,提出了合乘绕行分析方法。一方面,基于前景理论与演化博弈理论建立了绕行条件下乘客之间的合乘路径的博弈模型,分析了乘客路径博弈的稳定策略及其影响因素;另一方面,建立了绕行条件下司机的合乘策略模型,分析了司机合乘绕行策略的演化结果,并且引入投诉机制,进一步研究投诉机制下司机合乘策略的演化稳定策略及其影响因素,得出在合乘制度下投诉机制能够有效地防范司机拒载行为的结论。合乘绕行分析方法能够在有限理性条件下研究出现合乘绕行时的乘客及司机行为,从而进一步分析能够保证合乘绕行顺利进行的措施。本文从有限理性的新视角对出租车合乘行为展开研究。论文建立了合乘效果模型,提出了一种合乘效果分析方法;基于前景理论建立了乘客合乘行为决策模型,提出了一种乘客的合乘行为分析方法;将心理学理论引入行为建模,构建了司机收入公平模型,为合乘行为研究提供了新的思路;首次结合前景理论与演化博弈理论,提出了一种新的基于前景值的演化博弈方法,弥补了演化博弈模型中收益判断完全理性的缺陷;构建了乘客与司机的策略演化模型,分析了双方的稳定策略;并建立了合乘绕行条件下的乘客路径博弈模型及司机策略模型,专门对合乘绕行问题进行研究,为合乘问题做了完善性研究工作。论文对出租车合乘行为进行了深入分析,以新的视角开展了一系列有限理性条件下出租车合乘行为的基础理论研究,研究结论可为城市出租车合乘行为提供理论支持,对交通合乘政策的制定有重要指导意义。更多还原

【Abstract】 Taxi carpooling mode can alleviate city traffic congestion,save resources and reduce pollution.Many countries have implemented carpooling system,and achieved good results.Some cities in China are gradually implementing carpooling system to solve traffic problems.At present carpooling system in our country is in the initial stage,and there are many problems in the process of carpooling system implementation.The behaviors of passengers and drivers have great influences on carpooling implement.In reality,the phenomenon of a lot of passengers and drivers refusing to carpooling is very common,which will seriously hinder the implementation of carpooling system.Therefore,it is very necessary to study the behaviors of the passengers and drivers under the carpooling system.Considering bounded rationality of human,this paper studies the problem of taxi carpooling behavior with bounded rationality to make the research more close to the actual situation.In this paper,the researches on taxi carpooling behaviors are carried out based on prospect theory,evolutionary game theory and fair theory which are on the premise of bound rationality.The specific studies are as follows:(1)The research on taxi carpooling effects is carried out,and a simulation analysis method of taxi carpooling effects is put forward.Taxi carpooling model is built and simulation algorithm is designed based on analysis of passenger’s cost and driver’s income under different situations.Carpooling effects are analyzed by cost of passenger,income of the driver and capacity of taxi market.The influences of payment ratio and traffic congestion on carpooling effects are analyzed.Effectiveness of carpooling is test,and influences of carpooling effect are obtained.(2)The research on the carpooling behavior of passengers is carried out,and an analysis method of passengers’ carpooling decision is put forward.The model of passenger carpooling decision is built based on prospect theory with the decision factors of travel time,cost and comfort.Passenger capooling behavior and the influences on carpooling decision are analyzed with different reference points.The method proposed considers passengers psychology and uncertain traffic condition,embodies the characteristics of passengers’ bounded rationality,which is an analysis method of passengers’ carpooling decision with uncertain environment.(3)The research on the carpooling behavior of drivers is carried out,and an analysis method of drivers’ income and behavior is put forward.Drivers’ income fair model is established based on fair theory through the analysis of the drivers’ income and psychology with different conditions.The changes of average income,individual income and the influencing factors on income are analyzed.The model considers drivers’ behavior changes caused by their fair psychology,which provides a new thought for analyzing drivers’ carpooling behaviors with bounded rationality.(4)The research on the carpooling strategy of passengers and driver is carried out.The passengers’ strategy choices under different conditions are analyzed detailedly based on carpooling decision method proposed in the previous chapter.Prospect theory is introduced into evolutionary game theory,and evolutionary game model is improved.The capooling game model of the passengers and drivers based on prospect values is built.Then evolution analysis is done by replicated dynamic mechanism,and carpooling stable strategies of both sides are obtained.The method overcomes the problem of strategy gain rationality,and considers not only bounded rationality of gain judgment,but also dynamic adjustment of strategy.(5)The problem of carpooling detour is studied,and an analysis method of carpooling detour is put forward.On the one hand,carpooling path game model between passengers is built based on prospect theory and evolutionary game theory,and stable strategies and influence fators are analyzed.On the other hand,driver’s carpooling strategy model is established to analyze drivers’ evolution strategy.Furtherly,complaint mechanism is introduced,and drivers’ evolution strategies with the complaint mechanism are studied.It is found out that the complaint mechanism can prevent effectively drivers’ refusing behavior in the carpooling system.The method can analyze passengers and drivers’ s behavior under the condition of capooling detour with bounded rationality,and discusses the measures to make capooling detour possible.This paper studies the problem of taxi capooling behavior from a new perspective of bounded rationality.The carpooling effects model is built,and a method of taxi carpooling effects analysis is proposed;the passengers carpooling decision model is built based on prospect theory,and a method of passengers’ carpooling behavior analysis is proposed;psychology theory is introduced into behavior modeling,and the drivers’ income model is established,which provide a new thought for studying carpooling behavior;prospect theory and evolutionary game theory first are combined,a method of evolutionary game based on prospect values is proposed,which make up for full rationality shortcoming of earning judgment;strategy evolution model between passengers and behaviors is established to study stability strategy of behavior subjects;the carpooling path game model between passengers and driver’s carpooling strategy model are built under the condition of detour to study the problem of carpooling detour,which perfect carpooling researches.This paper thoroughly carries out fundamental theory research on taxi carpooling behavior from new perspective of bounded rationality.Research conclusions can provide theoretical support for carpooling behavior,which have important guiding significance for the development of traffic carpooling policy.更多还原