考虑鲁棒性的随机网络交通拥堵管理策略研究

发布时间：2018-04-16 01:11

本文选题：鲁棒性 + 随机交通网络　；参考：《南京大学》2015年硕士论文

【摘要】：现代社会中,交通拥堵问题已经成为一个严重的社会问题,城市交通网络的建设显得日益重要。相关学者和工作者提出了不少缓解交通拥堵问题的管理策略。然而由于现实交通网络中存在许多不确定因素,这些策略实际效果不尽如人意。因此,本文基于鲁棒性对随机网络中的交通拥堵管理策略进行了相关研究。针对最具代表性的拥挤收费策略和可交易路票策略,本文首先验证了确定性网络中两者的等价性,同时举例说明了随机交通网络中两者的差异。接着,对于可交易路票策略,本文假设随机网络中的所有状态是由一组不确定因素集合决定的,出行者选择所处状态下的最小出行成本路径。并用网络流量预测值与状态集的最小期望残量(ERM)来定义预测值的鲁棒性,同时通过光滑化梯度投影算法(SPG)对相关算例进行了求解。对于拥挤收费策略中的中心区域收费策略,将固定费率在一段时间内,能够有效控制中心区域车流量小于阀值的概率定义为为该费率下策略的可靠性,并由此来衡量策略对于交通拥堵控制的鲁棒性。此外,考虑了两种情景：(ⅰ)不完全信息网络,出行时间是随机的,出行者选取期望出行成本最小的路径；(ⅱ)完全信息网络,出行时间是Day-to-Day的模式,出行者选取当天出行成本最小的路径。基于这两种情景,分别进行了算例分析。结果表明：当可靠性越高时,每提高一个百分点的可靠性,所需要增收的收费费率也随之增高。针对此现象,本文进一步进行了灵敏度分析,并提出了相关的评价标准以平衡收费费率与可靠性。
[Abstract]:In modern society, traffic congestion has become a serious social problem, and the construction of urban traffic network is becoming more and more important.Scholars and workers have put forward a lot of management strategies to alleviate traffic congestion.However, due to the existence of many uncertain factors in the real traffic network, the actual effect of these strategies is not satisfactory.Therefore, based on robustness, traffic congestion management strategies in stochastic networks are studied in this paper.Aiming at the most representative congestion pricing strategy and tradable route ticket strategy, this paper first verifies the equivalence of the two in deterministic networks, and illustrates the differences between the two in stochastic traffic networks.Then, for tradable routing strategies, this paper assumes that all states in the stochastic network are determined by a set of uncertain factors, and the traveler chooses the minimum travel cost path in the state in which they are located.The robustness of the prediction value is defined by the network traffic prediction value and the minimum expected residual value of the state set. At the same time, the correlation example is solved by smoothing gradient projection algorithm (SPG).For the central area charging strategy in congestion charging strategy, the probability that the fixed rate can effectively control the traffic flow less than the threshold is defined as the reliability of the strategy under the fixed rate for a period of time.And then measure the robustness of the strategy to traffic congestion control.In addition, two scenarios: (I) incomplete information network are considered, the travel time is random, the traveler selects the path (II) complete information network with the least expected travel cost, and the travel time is the mode of Day-to-Day.Travelers choose the path with the lowest travel cost on that day.Based on these two scenarios, a numerical example is given.The results show that the higher the reliability is, the higher the charge rate is when the reliability is raised by one percentage point.In view of this phenomenon, the sensitivity analysis is further carried out, and the relevant evaluation criteria are put forward to balance the charge rate and reliability.
【学位授予单位】：南京大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491;O224

【参考文献】