Workshop: Advancing the Science of Transportation Demand Modeling; Berkeley, California; Spring 2017

交通运输对生活质量、经济、公平和环境有着巨大的影响。更好地规划、设计和运营我们的交通系统对社会至关重要。虽然交通需求模型被广泛用于指导交通基础设施和政策决策，但交通需求建模领域近年来因预测性能差而受到批评。预测不佳的原因之一是缺乏严格、科学的方法来开发和测试这些预测模型。与此同时，该领域正处于下一代模型的风口浪尖，这些模型是由即将上线的海量数据和数据分析方法的令人兴奋的发展所驱动的。来自计算机科学和物理学等学科的研究人员正在进入这一领域，但在很大程度上，他们与更传统的交通需求建模社区及其行为基础脱节。最后，不仅是新的数据和方法改变了建模，而且运输本身也在发生革命性的变化，包括清洁、连接、应用驱动和自主的运输模式/服务。鉴于这些限制、挑战和机遇，本次为期两天的研讨会旨在推进交通需求建模科学。研究人员和从业人员用来预测人类对交通政策的反应的现有知识和方法收集存在重大缺陷。运输需求建模研究界的操作实践倾向于促进竖井，并产生缓慢调整的方法范例。此外，它们没有提供激励或期望对运输需求预测系统进行严格测试、复制和验证。虽然社区试图解决这些缺点，但该领域只是在边缘进行调整，并没有导致重要的跨学科交流。研讨会将汇集来自不同学科、处于不同职业阶段的顶尖研究人员，从根本上重新思考交通需求建模领域，并就如何指导、实施、资助和评估研究提出新方法。与会者将包括来自传统和新兴领域的25-30名研究人员，他们专注于最先进的交通需求建模。目标是定义下一代运输需求模型，使其：(i)受到严格的科学测试，（ii）有效地整合来自不同学科的研究人员和想法，以及（iii）与迫在眉睫的运输转型（清洁，连接，自主等）密切相关。

Transportation has an immense impact on quality of life, economy, equity, and environment. Better planning, design, and operations of our transport system are of critical importance to society. While transport demand models are widely relied on to guide transport infrastructure and policy decisions, the transport demand modeling field has been criticized in recent times for poor forecast performance. Among the blame for the poor prediction is the lack of a rigorous, scientific approach of developing and testing of these forecasting models. At the same time, the field is on the cusp of the next generation of models driven by mega data coming online and exciting developments in data analysis methods. Researchers from disciplines such as Computer Science and Physics are entering the domain, but for the most part are siloed from the more traditional transport demand modeling community and its behavioral grounding. Finally, not only are new data and methods changing the modeling, but transport itself is being revolutionized with clean, connected, app-driven, and autonomous transport modes/services. In light of these limitations, challenges, and opportunities, this 2-day workshop aims to advance the science of transport demand modeling.There are major flaws in the existing knowledge and collection of methods that are used by researchers and practitioners to predict how humans respond to transport policies. The operating practices of the transport demand modeling research community tend to promote silos and produce methodological paradigms that are slow to adjust. Further, they do not provide incentives or expectations of rigorous testing, replication, and validation of the transport demand prediction system. While the community has attempted to address these shortcomings, the field is adjusting only at the margins and not leading to significant cross-disciplinary communication. The workshop will bring together leading researchers from different disciplines and at different stages in their careers to fundamentally rethink the field of transport demand modeling and propose new approaches on how research should be directed, conducted, funded, and evaluated. Attendance will consist of 25-30 researchers from traditional and new domains who focus on state-of-the-art transport demand modeling. The objective is to define the next generation of transport demand models such that it: (i) is subject to rigorous, scientific testing, (ii) effectively integrates researchers and ideas from different disciplines, and (iii) is germane to the looming transformation in transport (clean, connected, autonomous, etc.).