【作者】 黄一华；

【导师】 四兵锋；

【作者基本信息】 北京交通大学 ， 交通运输规划与管理， 2010， 硕士

【摘要】 随着城市轨道交通基础建设的不断发展,城市轨道交通已经成为特大、大型城市分担城市地面公共交通的重要组成部分。由于轨道交通网络规模的扩展,乘客到达目的地会存在多条路径,研究多路径情况下的乘客出行路径选择行为以及在此基础上的客流分配,具有重要的理论价值和现实意义。围绕轨道交通网络客流分配问题,本文主要进行了如下的研究工作：1、对北京市轨道交通乘客出行意愿调查数据进行整理、统计和分析,得到了不同类型的乘客在不同出行条件下对于路径选择的一般性规律和影响乘客路径选择的主要因素,并为模型的相关参数估计提供了数据依据。2、综合考虑影响乘客路径选择的定量因素,包括出行时间、换乘时间以及换乘次数,构造了路径广义费用模型。提出基于Logit关系的乘客路径选择概率模型。考虑北京市轨道交通网络拓扑的特点,采用改进Dijkstra算法和基于深度优先的图的遍历算法应用于有效路径搜索。通过对实例分析,验证本文所提出的模型及算法的有效性。3、考虑到乘客出行的舒适度问题,提出更加符合现实的路径广义费用模型。将Fisk随机用户均衡配流模型应用在城市轨道交通网络的配流问题中,利用MSA和MSWA(相继加权平均算法)两种算法进行求解。通过算例分析,比较了两种算法的收敛效果。更多还原

【Abstract】 As the Urban continuous development of rail transit infrastructure, urban rail transit has shared an important part of the city ground public transportation. The research of the passenger route choice behavior under multiple paths during traveling as well as the passenger flow distribution, has important theoretical value and practical significance.Around the problem of rail transit passenger flow assignment, the paper study as following:Firstly, statistic and analysis the survey data of Beijing rail transit. Obtain the main factors that impact on passenger route choice, and general rule for path selection of different types of passengers travel in different conditionsSecondly, considering the quantitative factors that influence the route choice behavior, improved the cost model by adding transfer time. Then, get the passenger route choice probability based on Logit model. Meanwhile, a modified Dijkstra algorithm and the depth first traversal algorithm for the model was given while searching effective routes in this paper. At last, an example is given to validate effectiveness of the model and algorithm.Finally, taking the issue of passenger comfort, the path cost should add the qualitative factor (congestion costs). Then, based on Fisk’s stochastic user equilibrium assignment model, use the MSA and the MSWA (successively weighted average algorithm) to solve the model. Examples are shown to compare the effect of two algorithms on convergence.更多还原