Multiple Stage Optimization of Stochastic Dynamic Transportation Networks

运输系统的决策过程是一个多阶段的问题，从长期的战略和战术规划决策到操作和实时/控制。在实时、业务和战术阶段经历的供需不确定性以及潜在控制决策的有效性会影响早期的战略决策。然而，在为网络优化做出战略决策时，这些未来的影响往往会被忽略。由于忽视了这些影响，规划者错过了开发更抗波动和能够充分利用信息潜力的网络的机会。这项研究将开发出计算交通系统随机性(和动力学)的方法，重点是将实时路线选择和控制等行为表示为资源。这将在前一阶段决策(如网络设计)的背景下使用，其中将调查新的方法以说明潜在的追索权。通过在总体战略、战术、操作和实时框架中查看系统，将探索阶段交互以及信息和不确定性在综合决策过程中的作用。在这种观点下，信息和控制的真正价值既可以从实时改进方面给予，也可以从改进对先前决定的追索权方面获得好处。除了研究问题外，还提出了几个基本的补充性教育问题。学生和实习运输工程师通常没有足够的技术背景来充分理解随机和动态系统的复杂行为，以及如何将实时数据的潜力用于实际的在线操作目的。这项研究将通过开发多阶段交通优化新课程的教材、教育模拟器和工具来明确解决这个问题，这些工具将成功地向学生传达与动态和随机交通系统的优化和控制相关的基本原理。通过进一步发展这一领域的基础研究，并将其纳入教育课程，未来的规划者将能够更好地考虑到存在的不可避免的不确定性，以及将在以后阶段提供的资源选择。反过来，这又允许创建一个更强大和更高效的交通网络。

The decision making process for transport systems is a multistage problem, ranging from long-term strategic and tactical planning decisions to operational and real-time/control. Uncertainty in supply and demand experienced in the real-time, operational and tactical stages as well as the effectiveness of potential control decisions impact earlier strategic decisions. However, these future impacts are often neglected when strategic decisions are made for the optimization of a network. By neglecting such impacts, planners are missing opportunities to develop networks that are more resistant to fluctuations and capable of exploiting the full potential of information.This research will develop methods for accounting for the transport system stochasticity (and dynamics) with a focus on representing actions such as real-time routing and control as recourse. This will then be used within the context of prior stage decisions (such as network design), where new methods will be investigated to account for the potential recourse. By viewing the system within the overall strategic, tactical, operational, and real-time framework, stage interactions will be explored as well as the role of information and uncertainty within the integrated decision process. The true value of information and control under such a view can be given both in terms of real-time improvement and with the benefits from improved recourse on prior decisions. In addition to the research problems, several fundamental complementary educational issues are raised in this context. Students and practicing transportation engineers do not generally have sufficient technical background to fully appreciate the complex behavior of stochastic and dynamic systems, and the know-how to exploit the potential of real-time data for actual online operational purposes. This research will explicitly address this problem by developing teaching material for a new course on multi-stage transportation optimization, educational simulators, and tools that will successfully convey the fundamental principles related to the optimization and control of dynamic and stochastic transportation systems to students. By further developing the fundamental research of this area, and integrating it into the education curriculum, future planners will be able to better account for the unavoidable uncertainties that exist as well as the recourse options that will be available in later stages. In turn, this allows for the creation of a more robust and efficient transportation network.