诱导信息条件下驾驶员路径选择决策研究

发布时间：2018-03-09 09:38

本文选题：行为强化理论　切入点：诱导信息　出处：《重庆交通大学》2015年硕士论文　论文类型：学位论文

【摘要】：本文主要围绕出行者的路径选择、诱导信息的发布以及模型博弈平衡进行了讨论与研究。首先,论文将行为强化理论应用到了驾驶员的路径选择行为中,提出了在有限理性条件下,驾驶员第k+1次的路径选择主要依赖于驾驶员自身第k次路径选择所获得的收益,即:自学习机制。其次,论文将驾驶员的行程时间感受作为驾驶员的路径选择收益,并将其时间感受划分成了三个模糊集,然后给出了各模糊集的隶属度函数。最后,论文在有限理性和以时间感受为模糊集的前提下,建立了相应的模型,并给出了各模型的求解算法与仿真验证。具体地来说,论文的主要研究工作内容如下:①将操作反射条件理论应用于驾驶员的车辆路径选择行为中,建立了基于有限理性模糊博弈无诱导信息条件下的车辆路径选择模型;证明了在模型的九种初始状态下,该博弈最终会取得平衡,并给出了九种初始状态下的博弈平衡结果。②建立了基于有限理性模糊博弈有诱导信息条件下的车辆路径选择模型,给出了该模型的求解算法,并采用具体算例对该模型进行了仿真验证。③对比分析了有诱导信息条件下和无诱导信息条件下的博弈平衡结果,结果显示:诱导信息的发布并不是一直有效的,诱导信息的有效性与参与博弈的局中人总数以及初始交通流的分布有密切关系;当参与博弈局中人总数小于或接近路网总通行能力时,发布诱导信息对减小局中人总行程时间是有利的,且路径L1的初始流量所占比例越接近C1/(C1+C2),诱导信息的有效性越差;诱导信息的有效性随路网总流量的增加而增加,当参与博弈局中人总数远大于路网总通行能力时,发布诱导信息能有效减小局中人总行程时间。④建立了累积自学习机制的车辆路径选择模型,给出了模型的求解算法,并采用具体算例进行了仿真验证,仿真结果表明:在累积自学习机制下,无诱导信息的车辆路径选择博弈结果与发布诱导信息的车辆路径选择博弈结果无显著差异。⑤分析了有诱导信息条件下的博弈模型参数ζ的变化对博弈平衡结果的影响。论文采用仿真验证的方法分析ζ的变化对博弈平衡结果的影响,仿真结果表明:当参与博弈车辆总数少于路网总通行能力时,ζ的变化对博弈平衡结果影响显著;当参与博弈车辆总数远大于路网总通行能力时,ζ的变化对博弈平衡结果无影响;在参与博弈车辆总数一定的情形下,初始状态接受诱导的车辆比例越大,ζ的变化对博弈平衡结果的影响越小。⑥最后论文分析讨论了博弈模型参数ζ的变化对诱导效果的影响。结果表明:ζ能影响局中人总行程时间T总,它对T总的影响显著性与博弈的初始接受诱导的车辆比例及博弈局中人总数相关;当局中人总数小于(或接近)路网总通行能力时,ζ对T总的影响主要与路网流量的初始接受诱导的车辆比例有关;当局中人总数远大于路网总通行能力时,ζ对T总影响不显著。
[Abstract]:This paper mainly focuses on the travelerroute choice, information release and induced model of game equilibrium are discussed and studied. Firstly, the behavior reinforcement theory is applied to the driver's route choice behavior, proposed in the limited rationality, driver path selection in k+1 mainly bases on the driver's own K path selection of gains, namely: self-learning mechanism. Secondly, the travel time of driver's feelings as the driver's route choice and the time to feel benefits, divided into three fuzzy sets, and then gives the fuzzy set membership function. Finally, in the premise of limited rationality and time to feel for fuzzy sets, established the corresponding model, and gives the algorithm and Simulation of the validation of the model. Specifically, the main contents are as follows: 1. The operation of reflection The vehicle routing application conditions to the driver's choice behavior theory, established the finite rational fuzzy game by vehicle route choice model based on information conditions; proved that nine kinds of initial state model, the game will eventually achieve a balance, and the balance state of the game results was nine at the beginning of the establishment are given. The limited rationality of fuzzy game induced by vehicle path under the condition of information selection model based on the algorithm of the model is given, and the example for the simulation of the model. The comparison and analysis of the induction information and by game equilibrium results, under the conditions of information showed that the induction and release the information is not always effective, there is a close relationship between the effectiveness of the guidance information bureau and in the total number of people involved in the game and the initial traffic flow distribution; when the total number of players in the game Less than or close to the total network capacity, issued guidance information is favorable to the reduction in the total travel time of most people, and the initial flow path L1 proportion is close to C1/ (C1+C2), the effectiveness of the guidance information is poor; the effectiveness of the guidance information increases with the increase of the total network traffic, when in the total number of game players is far greater than the total network capacity, can effectively reduce the induced release of small players total travel time information. The model utilizes the vehicle routing self-learning mechanism selection model, gives the algorithm to solve the model, and the example was simulated, simulation results show that the accumulation of self learning mechanism, no vehicle route guidance information selection results of the game with the release of vehicle route guidance information selection game results have no significant difference. 5. The induced game model parameters under the condition of the change of zeta information on Effect of game equilibrium results. The method changes the simulation analysis on the game of the zeta balance results, simulation results show that when the total number of vehicles in the game less than the total network capacity, the impact of change on the balance of the game. In the game when the total number of vehicles significantly; far greater than the total network capacity, zeta the change has no effect on game equilibrium results; the total number of vehicles in the game in certain circumstances, the initial state of the vehicle induced greater proportion of acceptance, the change of zeta balance results is small. The paper finally analyzes the influences of model parameters on the change of the game zeta induction effects discussed. The results show that the zeta can the people in the game affect the total travel time of T, its impact on the overall T initial significant and game total vehicle proportion and accept by game players; the total number of people in authority (or close to less than ) the total network capacity, the main effect of T zeta total with the initial network flow induced by the vehicle accept proportion; total authorities people far greater than the total network capacity, there was no significant difference on the total effect of zeta T.

【学位授予单位】：重庆交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U495

【参考文献】