CAREER: The Evolution of Transportation Networks: Empirical Research and Agent-Based Models

摘要摘要：交通网络的演化：实证研究与基于主体的模型本研究致力于从理论和实证层面理解交通网络的演化增长过程，认识到供给和需求之间的相互依赖关系。要检查的主要问题包括：为什么网络会扩展和收缩？网络是否会自组织成层次结构？道路(路线)是网络的新兴属性吗？哪些投资规则预测了网络改进的顺序和位置？与提供新设施(新连接)相比，何时扩展现有设施(同一路段上的更多车道)？如何改进交通规划，以利用对网络演变的新理解？假设简单、可测量的因素(如交通增长率、运量与运力比率，以及与相邻上下游链路的比较)可以解释许多由此产生的决定。这项调查将审查明尼阿波利斯和圣保罗这两个孪生城市的资本改善项目和决策的时间序列，将它们与网络结构特征联系起来。这些经验模型将被嵌入到基于代理的模型中，以复制网络增长的过程。本研究详细说明并估计了解释旅行行为、网络成本和收入以及投资决策的组件模型。然后将组件集成到仿真模型中并进行测试。无论是对单个组件模型的估计，还是将它们集成到网络增长(和下降)的模拟中，都将增加我们对网络演化过程的有限理解。这一新的认识将对交通规划实践产生更广泛的影响，并最终对城市和区域的形态产生影响。在决策中纳入明确的网络外部性测量将带来更好的计划、网络路由决策和实施策略。理解和说明在一个时间点上的决定如何影响未来的选择，应该有助于指导希望塑造未来的规划者和决策者。将评估渐进式变化的长期后果。这将有助于决策者评估扩大现有设施或路线、建设新的通行权或提供新服务的影响。这种对长期网络动态的更好理解将导致更好地规划和设计公路网，以利用网络外部性，并最大限度地增加决策者的未来选择。了解当前的决策如何排除或创造未来的机会应该会改善决策。该研究的主要教育目标是：1)将交通网络的演变整合到城市交通规划、交通经济学、交通与土地使用的课程中；2)开发一个关于交通政策的新生研讨会，让明尼苏达大学的本科生通过本科生研究机会计划参与进来；3)让来自丰迪拉克部落的美国原住民高中生参与交通运输。在课程中使用模拟模型将为学生提供一个独特的工具，帮助他们理解策略决策对网络发展的影响。

AbstractAbstract - CAREER: The Evolution of Transportation Networks: Empirical Research and Agent-Based ModelsThis research endeavors to understand the evolutionary growth process of transportation networks at a theoretical and empirical level, recognizing the inter-dependence of supply and demand. Key questions to be examined include: Why do networks expand and contract? Do networks self-organize into hierarchies? Are roads (routes) an emergent property of networks? What investment rules predict the sequence and location of network improvements? When are already existing facilities expanded (more lanes on the same link) as opposed to new facilities being provided (a new link)? How can transportation planning be improved to take advantage of a new understanding of network evolution?It is hypothesized that simple, measurable factors (such as traffic growth rates, volume to capacity ratios, and comparison with adjacent upstream and downstream links) explain many of the resulting decisions. This investigation will examine a time series of capital improvement projects and decisions for the Twin Cities of Minneapolis and St. Paul, relating them to network structure characteristics. These empirical models will be embedded in agent based models to replicate the process of network growth.This research specifies and estimates component models explaining travel behavior, network costs and revenue, and investment decisions. Then the components will be integrated into the simulation model and tested. Both the estimation of individual component models and their integration into a simulation of network growth (and decline) will increase our limited understanding of network evolution processes. This new understanding will have broader impacts on transportation planning practice, and ultimately on the shape of cities and regions. Incorporating explicit measures of network externalities in decision making will lead to better plans, network routing decisions, and implementation strategies. Understanding and illustrating how decisions in one point of time affect future choices should help guide planners and decision-makers desiring to shape the future. The long-term consequences of incremental changes will be assessed. This will help decision-makers assess the effects of expanding existing facilities or routes, or building in new rights-of-way or offering new services. This improved understanding of long term network dynamics would lead to better planning and design of road networks to exploit network externalities and maximize future choice for decision makers. The knowledge of how current decisions foreclose or create future opportunities should improve decision-making.The primary educational objectives of the research are to 1) integrate transportation network evolution into courses on Urban Transportation Planning, Transportation Economics, and Transportation and Land Use, 2) develop a Freshman Seminar on Transportation Policy to involve University of Minnesota undergraduates through the Undergraduate Research Opportunities Program, and 3) involve Native American High School Students from the Fond du Lac tribe in Transportation. Use of the simulation model in the courses will provide the students a unique tool for understanding the implications of policy decisions on network evolution.