基于分层Logit的多方式随机用户均衡分配模型  

Multimodal stochastic user equilibrium model based on nested Logit

作　　者：刘诗序[1] 池其源 阎昊[1] 郭丽丹 唐颖诺 严海[3] LIU Shi-xu;CHI Qi-yuan;YAN Hao;GUO Li-dan;TANG Ying-nuo;YAN Hai(School of Civil Engineering,Fuzhou University,Fuzhou 350116,Fujian,China;Fujian Provincial Expressway Technology Consulting Co.,Ltd.,Fuzhou 350108,Fujian,China;Beijing Collaborative Innovation Center for Metropolitan Transportation,Beijing University of Technology,Beijing 100124,China)

机构地区：[1]福州大学土木工程学院,福建福州350116 [2]福建省高速技术咨询有限公司,福建福州350108 [3]北京工业大学北京城市交通协同创新中心,北京100124

出　　处：《长安大学学报（自然科学版）》2018年第5期114-122,共9页Journal of Chang’an University(Natural Science Edition)

基　　金：国家自然科学基金项目(51308126;51308018);福建省教育厅中青年教师教育科研项目(JAT170065);北京市属高等学校人才强教计划项目

摘　　要：为了完善现有的交通分配模型,综合考虑出行者的方式选择与路径选择行为,建立了基于分层Logit的多方式随机用户均衡分配模型。首先,在多方式交通网络中,根据出行者出行决策的2个步骤,即首先选择合适的交通出行方式、然后选择该方式交通网络中的出行路径,采用分层Logit模型描述出行者的交通方式和路径联合选择行为。在建立的分层Logit模型中,出行效用与流量分布相互制约,出行效用受路网流量影响,而路网流量由效用所决定,最终要达到平衡。接着,构造了一个数学规划模型来求平衡解,证明数学规划模型的最优解与分层Logit模型的流量解等价,并进一步证明等价数学规划模型的目标函数是凸的,而约束集也是凸的,因此等价数学规划模型是一个凸规划问题,模型的解是唯一的。随后,给出了求解提出的基于分层Logit的多方式随机用户均衡分配模型改进的方向搜索算法步骤。最后,以含公交车和小汽车2种方式的方格形路网为算例,分析公交票价、出行者的时间价值和分层Logit模型的层间比例参数的变化对交通分配结果的影响。结果表明:公交车票价对流量分配的影响不大;公交需求量随出行者时间价值的增加而下降;当出行者的时间价值较低时,公交需求量随分层Logit模型的层间比例参数增加而增加,当出行者的时间价值较高时,公交需求量随分层Logit模型的层间比例参数的增加而减小。In order to improve the current urban traffic assignment models,a multimodal stochastic user equilibrium model based on nested Logit was proposed,considering the behaviors of mode choice and route choice.Firstly,in a multimodal traffic network,a nested Logit model was proposed to describe the behavior of the combined traffic mode and route choice,according to the two steps of a traveler’s travel decision（he or she chooses an appropriate traffic mode and then chooses a route in the traffic network）.In the proposed nested Logit model,the utility and traffic flow distribution were mutually restricted,the utility of the travel was influenced by thenetwork traffic flow,and the traffic flow was determined by the utility,thus leading to equilibrium.Therefore,a mathematical programming model for finding the equilibrium solution was developed in this study.It was proved that the optimal solution of the mathematical programming model and the flow solution of the nested Logit model were equivalent.It was further proved that both the objective function of the equivalent mathematical programming model and the constraint set were convex.Therefore,the equivalent mathematical programming model was a convex programming problem,and the solution of the model was unique.Subsequently,the step of improved the direction search algorithm was given to solve the proposed multimodal stochastic user equilibrium assignment model based on nested Logit.Finally,a case study of a grid network with two modes,namely bus and car,was carried out to analyze the impact of bus fare,traveler’s time value,and the interlayer ratio parameter of the nested Logit model on traffic assignment results.The results show that the changes in bus fare have little effect on the traffic assignment results,and that bus demand decreases with increase in travelers’time value.When the travelers’time value is low,the bus demand increases with increase in the interlayer ratio parameter of the nested Logit model,whereas when the travelers’time value is high,the

关 键 词：交通工程 随机用户均衡 LOGIT模型 多方式网络 路径选择行为 

分 类 号：U491[交通运输工程—交通运输规划与管理]