组合出行模式下交通流分配模型及其求解研究

发布时间：2018-05-04 18:37

本文选题：组合出行 + 交通流分配　；参考：《兰州交通大学》2014年硕士论文

【摘要】：城市化进程的迅猛发展，对城市交通提出了新的要求，各种环保、节能、高效的交通方式被应用，我们的生活已经离不开BRT、地铁等交通方式。在城市规模越来越大的同时，出行者的出行模式将从单一的交通方式转变为多种交通方式联运，这种单一方式和换乘方式并用的出行方式成为组合出行。参阅大量相关文献可知，组合出行下的交通流分配将成为交通运输规划与管理的核心。静态交通流分配和动态交通流分配是两个主要研究方向，二者最根本区别是路段的阻抗是否为一个时间函数。前者的路阻不变，在实际情况中不可能实现，因此动态交通流分配是研究的重点。但对静态交通流分配的研究也不容忽视，静态研究重点是宏观的交通现象，对分析未来交通的发展趋势有重要作用。本文从以下方面对组合出行模式下的交通流分配展开研究：首先，对超级网络进行简要介绍，建立以兰州市安宁区兰州交通大学附近实际路网为例的超级网络，以此作为本文研究的基础。对超级网络的定义、构成和拓扑结构分析，使用图的遍历法找出所有路径，然后根据有效超路径的定义，在所有路径中选择换乘次数最多两次的路径，并对这个过程进行算法研究。其次，分别从静态和动态的角度对交通流分配进行研究。假设出行者的出行模式和路径都是随机的，，提出模型约束条件，使用logit模型表达均衡条件，建立基于弹性需求满足随机用户均衡条件的logit模型，因为对logit模型直接求解较为复杂，所以本文构造了与logit模型等价的变分不等式模型，证明其等价性和最优解的存在性；采用基于权重平均法设计的连续权重平均法，对等价变分不等式模型进行求解，从而获得基于弹性需求满足随机均衡条件的logit模型的解。根据对组合出行模式下静态交通流分配模型的研究，分析约束条件，构建动态交通流分配模型。此模型也需满足随机用户均衡，同时考虑出行路径和出行时间的选择。用logit模型对出行行为进行描述，并用等价的变分不等式模型求解，求解的算法使用启发式迭代算法。然后，列举出一个小型路网，用静态交通流分配模型和动态交通流分配模型对该路网分别进行分析，研究模型和算法的可靠性和高效性，并对动态交通流分配模型的一些参数进行分析并进行说明，阐述交通流的OD量和费用等发生的变化。以兰州市安宁区兰州交通大学附近实际路网为例，作为中型路网案例，构建较为复杂的超级网络，对网络的路径、节点等组成部分进行赋值，使用动态交通流分配模型对该中型路网进行计算。最后，采用静态交通流分配模型对小型路网分析，得出结论，出行者对路网信息了解的越多，路网的不确定性减小，出行者更容易做出更加明确和明智的选择。采用动态交通流分配模型应用于两个算例分析得出，动态交通流分配模型能够较为真实的反映出行者的偏好对出行行为的影响，分析了参数之间变化情况对交通需求和费用的影响，能够同时满足对选择出行路径和出行时间的要求，求解算法可靠且高效，可以应用于中小型路网的研究，具有一定的实用性。
[Abstract]:With the rapid development of urbanization process, new requirements are put forward for urban traffic. All kinds of environmental protection, energy saving and efficient transportation are applied. Our lives can not be separated from BRT, subway and other modes of transportation. As the city is growing larger, the traveler's travel mode will be transformed from a single traffic mode to a variety of transportation mode. This single mode and the way of transfer are combined trips. In a large number of related literature, the traffic flow allocation under combined travel will be the core of transportation planning and management. Static traffic flow distribution and dynamic traffic flow allocation are two main research directions, and the most fundamental difference between the two is the impedance of the section. It is not a time function. The road resistance of the former can not be realized in the actual situation, so the dynamic traffic flow distribution is the focus of the study. But the research on the static traffic flow assignment can not be ignored. The static research focuses on the macroscopic traffic phenomenon and plays an important role in the analysis of the trend of the future traffic development. Research on traffic assignment under combined travel mode:
Firstly, the super network is briefly introduced, and a super network is set up with the actual road network near LanZhou JiaoTong University in Lanzhou Anning District, which is the basis of this study. The definition, composition and topology analysis of super network are used to find the path by the traversal method of the graph, and then, according to the definition of the effective superpath, all of them are defined. In the path, select the maximum number of transfers two times, and study the algorithm.
Secondly, we study the traffic flow distribution from the static and dynamic perspectives, assuming that the traveler's travel mode and path are random, and the model constraints are proposed, the logit model is used to express the equilibrium conditions, and the logit model based on the elastic demand satisfies the stochastic user equilibrium condition is established, because the logit model is solved directly. Therefore, this paper constructs a variational inequality model equivalent to the logit model, proves its equivalence and the existence of the optimal solution, and uses the weight averaging method to solve the equivalent variational inequality model by means of the weight averaging method, thus obtaining the solution of the logit model based on the elastic demand satisfying the stochastic equilibrium condition. The static traffic flow allocation model under the combined travel mode is studied and the constraints are analyzed and the dynamic traffic flow distribution model is constructed. This model also needs to satisfy the random user equilibrium, while considering the choice of travel path and travel time. The logit model is used to describe the travel behavior, and the solution is solved with the equivalent variational inequality model. The algorithm uses a heuristic iterative algorithm.
Then, a small road network is enumerated, the static traffic flow distribution model and dynamic traffic flow distribution model are used to analyze the road network respectively. The reliability and efficiency of the model and algorithm are studied, and some parameters of the dynamic traffic flow distribution model are analyzed and explained, and the changes of the traffic flow OD and cost are explained. Taking the actual road network near LanZhou JiaoTong University in Lanzhou Anning District as an example, as a medium-sized network case, a more complex super network is constructed, the path of the network, nodes and other components are assigned, and the dynamic traffic flow distribution model is used to calculate the medium network.
Finally, the static traffic flow distribution model is used to analyze the small road network. The conclusion is that the more information of the road network information, the less uncertainty of the road network, the more clear and wise choice of the traveler. The dynamic traffic flow allocation model is applied to two examples, and the dynamic traffic flow distribution model can be obtained. It is more realistic to reflect the influence of traveler's preference on travel behavior, analyze the influence of change between parameters on traffic demand and cost, and meet the requirement of travel route and time. The algorithm is reliable and efficient. It can be applied to the research of medium and small road network, and it has certain practicability.

【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.12

【参考文献】