行人过街过程中交通主体行为演化研究

发布时间：2018-03-30 04:05

本文选题：过街行人　切入点：机动车驾驶员　出处：《北京交通大学》2015年博士论文

【摘要】：由行人、机动车、非机动车等交通主体组成的混合交通对城市交通的影响及作用越来越明显。交通主体行为已经不再是影响城市交通运行状态的干扰因素，而是决定城市交通运行状态的主导因素。开展对交通主体行为的研究具有十分重要的理论和现实意义。己有的研究成果主要集中在交通主体行为特征、交通主体行为模拟、交通主体行为心理成因等方面，对交通主体行为的演化过程、演化规律研究较少。对于这一演化过程和演化规律，传统的交通流模型己经不能充分的描述和解释，迫切需要一种能够解释演化过程和演化规律的理论。本文以复杂自适应理论、演化经济学和演化博弈论为基础，将行人过街过程的交通主体行为作为研究对象：从演化的角度研究交通主体行为演化规律。在对交通土体行为进行系统解析的基础上，首先构建基于空间演化博弈的行人过街行为演化模型，研究行人过街行为演化规律：其次构建基于时间延误函数的过街行人和机动车驾驶员行为演化阵弈模型，研究机动车驾驶员和过街行人行为演化规律：最后构建行人过街多群体行为非对称动态熨制模型：研究过街行人、机动车驾驶员和交通管理人员的行为演化规律。木文具体研究内容与结论如下：在描述交通主体行为系统的基础土，分析交通主体行为具有复杂自适应和动态演化特征．从行为惯例、行为演变、社会文化选择、行为保持方面分析交通主体行为的演化机制，从强制演化、自我学习演化、强制演化与自我学习演化交叉融合方面分析交通主体行为的演化方式。在空间演化博弈的行人行为演化模型中，将重复博弈与Agent相结合，构造过街行人行为决策函数，提出过街行人自我学习与相互学习的演化机制．研究各类人群行为演化规律，过街设施方便性、交通管理、交通处罚与违章行为的关系发现：（1）违章处罚具有闭值，违章处罚的额度比执法的严格性具有更大的效力：（2）过街设施极其方便和极其不方便时，过街行人对过街设施的方便性不敏感：（3)过街设施方便性较差时，交通管理对过街行人的违章行为不起作用；（4)遵守规则的过街时间、过街的平均时间和冒险违章行人比例三者之间两两具有正相关关系；冒险违章行人比例与过街失败时间具有负相关关系．在时间延误函数的过街行人和机动车驾驶员行为演化模型中，引入机动车驾驶员和过街行人的时间延误函数，构建能够刻画信号交叉口过街行人和机动车驾驶员穿越行为的长期演化模型；并在此基础上研究机动车驾驶员和过街行人各自收益、损失的变化对长期演化结果的影响。通过研究发现：信号交叉口的平均安全状态、调控标准；过街行人和机动车驾驶员长期演化的目标是相互礼让；通过调控行人流、机动车流最、机动车速度和行人过街速度等决策参数：可以优化过街行人和村L动车驾驶员行为的长期演化路径，诱导过街行人和机动车驾驶员相互礼让，提高过街效率：在行人过街多群体行为非对称动态复制模型中，从过街行人、机动车驾驶员和交通黄理人员的风险认知、行为选择入手，探讨不同情境下交通管理人员、过街行人、机动车驾驶员行为演化的一般规律。分析演化稳定策略、混合策略纳什均衡．仿真行为演化规律发现：(1)系统状态收敛到过街行人、机动车驾驶员违章，交通管理人员认真管理，机动车驾驶员和过街行人为了获得更大的利益，采取了对自己最有利的策略，，结果是均获得次优收益，未实现社会福利和自身收益最大化：(2)过街行人、机动车驾驶员和交通管理人员行为演化其有明显的路径依赖特征：(3过街行人、机动车驾驶员和交通管理人员行为演化策略的均衡不仅仅取决于自身收益，还受到其他两群体收益的影响：(4)交通主体行为影响因子包括利导因子和限制因子，交通主体行为演化过程是在两种因子的双重作用下完成。本文的主要创新点如下：(1)摒弃传统的交通主体行为假设，以演化为研究范式，将交通主体行为“嵌入”到交通环境中；从交通情景、行为适应与演化、交通情景的涌现与强化的视角对行人过街过程中交通主体行为的长期演化进行研究。(2)在交通主体行为系统演化解析的基础上，基于“学习人”假设，建立行人过街行为的空间演化博弈模型。(3)从过街行人、机动车驾驶员和交通管理人员的风险认知、行为选择入手，建立基于时间延误函数的行人与机动车驾驶员行为演化模型、多群体行为非对称演化模型，分析交通主体之间的互动关系及行为演化规律。
[Abstract]:This paper studies the evolution law of traffic subject behavior based on complex adaptive theory , evolutionary economics and evolutionary game theory . This paper analyzes the evolution mechanism of traffic subject behavior from the aspects of behavior practice , behavior evolution , social culture selection and behavior maintenance . In the pedestrian behavior evolution model of spatial evolution game , the repeated game and agent are combined , the pedestrian behavior decision function is constructed , and the evolution mechanism of pedestrian ' s self - learning and mutual learning is proposed . It is found that : ( 1 ) It has a closed value , and the amount of disciplinary punishment is more effective than the strictness of law enforcement : ( 2 ) The crossing facilities are extremely convenient and extremely inconvenient , the pedestrian crossing the street is not sensitive to the convenience of street facilities : ( 3 ) When the convenience of the street facilities is poor , traffic management has no effect on the behavior of pedestrian crossing over the street ;
( 4 ) The crossing time of the rule , the average time of crossing the street and the pedestrian proportion of adventure violation have positive correlation ;
The pedestrian proportion of risk - taking violation has a negative correlation with over - street failure time . In the pedestrian and motor vehicle driver behavior evolution model of the time delay function , the time delay function of the driver of the motor vehicle and the pedestrian crossing the street is introduced , and a long - term evolution model can be constructed which can depict the pedestrian crossing behavior of the pedestrian and the driver of the motor vehicle at the signalized intersection .
On the basis of this , the effects of the individual income and loss on the long - term evolution of the drivers of the motor vehicle and over the street are studied .
The goal of long - term evolution of pedestrian and motor vehicles across the street is mutual comity ;
By controlling the decision parameters such as pedestrian flow , motor vehicle flow , motor vehicle speed and pedestrian crossing speed , it is possible to optimize the long - term evolution path of driver ' s behavior of pedestrian and village L . The simulation behavior evolution law finds that : ( 1 ) The system state converges to the pedestrian in the street , the driver of the motor vehicle , the traffic management personnel seriously manage , the driver of the motor vehicle and the pedestrian in the street are seriously managed . The equilibrium of the behavior evolution strategy of the pedestrian , the motor vehicle and the traffic management personnel has obvious path dependence .
This paper studies the long - term evolution of traffic subject behavior in the process of pedestrian crossing from the perspective of traffic situation , behavior adaptation , evolution and traffic situation . ( 2 ) Based on the analysis of the evolution of traffic subject behavior system , the spatial evolution game model of pedestrian crossing behavior is established based on the hypothesis of " learning man " .

【学位授予单位】：北京交通大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：U491

【参考文献】