灰色马尔可夫模型在交通事故预测中的应用

采用传统灰色GM(1,1)模型预测道路交通事故这类随机性、波动性较大的数据,存在拟合较差、精度不足等问题。运用马尔可夫模型对预测结果进行优化,并建立灰色马尔可夫预测模型。将该模型应用于1998~2007年全国道路交通事故数据来预测2008年事故次数,结果表明采用该模型预测精度达到96.9%,明显优于单独使用GM(1,1)模型的预测结果。     

郑州市公交价格需求弹性研究

本文结合郑州市公交价格调整实际,研究并估算郑州市BRT、空调公交、全部地面公交的价格需求弹性,研究表明:郑州市地面公交价格需求弹性总体较小.最后分析了其他因素时郑州市公交需求的影响,并在规划、管理、服务等方面提出了城市公交发展的部分建议.     

基于非等时序灰色模型的路基沉降预测

结合高速公路路基施工沉降观测数据,讨论了灰色系统理论和双曲线模型在公路路基沉降预测中的应用,并对等间隔的灰色模型GM(1,1)进行了改进,建立了任意时间间隔的非等时序改进灰色模型。通过具体工程实践,给出了灰色模型和双曲线模型对公路路基沉降量预测结果与实测结果的比较,结果表明灰色模型的预测沉降量与实际沉降量更接近,精度更高,更能满足工程需要。     

Robust optimization for emergency logistics planning: Risk mitigation in humanitarian relief supply chains

This paper proposes a methodology to generate a robust logistics plan that can mitigate demand uncertainty in humanitarian relief supply chains. More specifically, we apply robust optimization (RO) for dynamically assigning emergency response and evacuation traffic flow problems with time dependent demand uncertainty. This paper studies a Cell Transmission Model (CTM) based system optimum dynamic traffic assignment model. We adopt a min–max criterion and apply an extension of the RO method adjusted to dynamic optimization problems, an affinely adjustable robust counterpart (AARC) approach. Simulation experiments show that the AARC solution provides excellent results when compared to deterministic solution and sampling based stochastic programming solution. General insights of RO and transportation that may have wider applicability in humanitarian relief supply chains are provided.     

Robust Wardrop’s user equilibrium assignment under stochastic demand and supply: Expected residual minimization approach

Various models of traffic assignment under stochastic environment have been proposed recently, mainly by assuming different travelers’ behavior against uncertainties. This paper focuses on the expected residual minimization (ERM) model to provide a robust traffic assignment with an emphasis on the planner’s perspective. The model is further extended to obtain a stochastic prediction of the traffic volumes by the technique of path choice approach. We show theoretically the existence and the robustness of the ERM solution. In addition, we employ an improved solution algorithm for solving the ERM model. Numerical experiments are carried out to illustrate the characteristics of the proposed model, by comparing with other existing models.     

Quantifying the benefit of responsive pricing and travel information in the stochastic congestion pricing problem

This paper is concerned with roadway pricing amidst the uncertainty which characterizes long-term transportation planning. Uncertainty is considered both on the supply-side (e.g., the effect of incidents on habitual route choice behavior) and on the demand-side (e.g., due to prediction errors in demand forecasting). The framework developed in this paper also allows the benefits of real-time travel information to be compared directly against the benefits of responsive pricing, allowing planning agencies to identify the value of these policy options or contract terms in publicly-operated toll roads. Specifically, six scenarios reflect different combinations of policy options, and correspond to different solution methods for optimal tolls. Demonstrations are provided on both the Sioux falls and Anaheim networks. Results indicate that providing information to drivers implemented alongside responsive tolling may reduce expected total system travel time by over 9%, though more than 8% of the improvement is due to providing information, with the remaining 1% improvement gained from responsive tolling.     

The use of the bicycle

The rediscovery of the bicycle by the public, by politicians and by professional urban transportation planners as a mode of transport which is perfectly in harmony with the goals of environmental protection, energy saving and personal fitness has stimulated this empirical study on the actual use of the bicycle by various population groups for obligatory and discretionary trip purposes. The influence on bicycle usage of such factors as age, education, car availability, residential density and town size, topography and time of year is analysed in this paper for selected population groups. For housewives from motorized households logit‐models were designed and calibrated to model their modal choice for shopping trips with special references to the bicycle. From the empirical results, the groups with the largest potentials for cycling are identified and the extent to which the potentials could be activated by specific policies is discussed. The research is based on a large sample held to be representative for the Federal Republic of Germany in 1976 and is supplemented by more recent surveys in selected German cities conducted by SOCIALDATA Munich.     

基于改进的灰色BP神经网络的区域物流成本预测

区域物流成本的测量与分析是考察一个地区物流水平的关键问题,合理地计算和预测物流成本可以使各级政府对物流发展采取有效的宏观调控措施。本文提出了适合中国地区的物流成本的较合理的测算方法,以吉林省为例进行了实证分析,并采用改进的灰色BP神经网络模型对吉林省的物流成本进行预测,取得了满意的预测结果。研究表明控制物流成本仍然是振兴区域经济,提高经济效益的源泉。     

交通量预测的神经网络集成方法

首次将神经网络集成技术引入交通量预测。神经网络集成通过训练多个神经网络并将各网络输出进行合成,可显著提高学习系统的泛化能力。在Boosting和Bagging集成方法的基础上,提出基于分治策略的神经网络集成方法,并且讨论了网络权重分配算法。使用上述三种神经网络集成预测模型,对苏州某交叉口实时交通量进行预测,预测结果比较理想,优于单一神经网络预测方法。实验表明,神经网络集成用于交通量预测是有效可行的。     

Recent developments in Flexible Transport Services

This paper focuses on recent experiences in the evolution of Flexible Transport Services (FTS). It describes how previous manifestations of FTS provided dedicated solutions for special groups, often in parallel to the conventional public transport network, but highlights that in numerous countries across Europe there is now a discernable trend towards open access FTS providing services for the whole community and often feeding the conventional public transport network to provide wider access and opportunities. This is as a result of policies intended to mainstream services for people with disabilities and moves to tackle social exclusion amongst other groups while at the same time reducing the very high costs per trip associated with STS.To meet these policies, and the higher expectations and different travel behaviours of citizens to-day, it is suggested that FTS will need to radically reposition itself in terms of scale, so that there is mass coverage and ability to function as a full transport mode. This will require breakthroughs in concepts; business models; organisational and operational models; and in supporting technologies. A proposed solution to this is the Flexible Agency for Collective Mobility Services (FAMS) which provides an organisational structure and business model for FTS that incorporates the required supporting technologies.