Dagstuhl-Seminar 27111
Dynamic Traffic Models in Transportation Science
( 14. Mar – 19. Mar, 2027 )
Permalink
Bitte benutzen Sie folgende Kurz-Url zum Verlinken dieser Seite:
https://www.dagstuhl.de/27111
Organisatoren
- Hong K. Lo (The Hong Kong Univ. of Science & Technology, HK)
- Carolina Osorio (HEC Montréal, CA)
- Nicolas Stier-Moses (Meta - Menlo Park, US)
- Laura Vargas Koch (RWTH Aachen, DE)
Kontakt
- Marsha Kleinbauer (für wissenschaftliche Fragen)
- Simone Schilke (für administrative Fragen)
Accurately modeling and predicting urban traffic is crucial for the strategic design and operations of transportation networks. Cities and companies worldwide use traffic models and optimization frameworks to inform a variety of decisions, such as developing novel services and offerings based on autonomous and/or electric vehicle fleets. This becomes increasingly important in the era of real-time information. One fundamental issue that needs to be addressed is that when realistic traffic flow characteristics such as physical queuing, spillback, and shock waves are considered, large-scale problems become highly complex and difficult to solve. In this seminar, the fifth in the series, our objective is to bring together leading academics and private-sector researchers from the traffic simulation, algorithmic game theory, and dynamic traffic assignment (DTA) areas to jointly address this issue from the combined point of view. Especially, we aim to bring together: (i) traffic scientists who formulate realistic urban mobility problems and rely on detailed urban traffic simulations to address them, with (ii) theoreticians from mathematics and computer science who aim to formally understand the underlying properties of the models, based on theoretical proofs.
In this Dagstuhl Seminar fundamental questions connecting DTA and characteristic traffic flows will be addressed. We aim to explore the following topics that operate between approaches based on theory and simulation:
- Flows over Time and Dynamic Traffic Assignment is the long-standing congregating topic that brought the community of this seminar series together.
- Learning approaches can be used to efficiently compute dynamic equilibria and their approximations. Learning requires real-time information. This interacts with privacy considerations which require constraints on what data representations are possible.
- Mechanism design can produce socially-desirable dynamic equilibria using classic tools, such as tolls, but also information design. This requires anticipating how traffic reacts to the provision of information.
- Partial differential equation formulations provide accurate solutions but they are hard-to-analyze because they are based on physical propagation models.
Creative Commons BY 4.0
Hong K. Lo, Carolina Osorio, Nicolas Stier-Moses, and Laura Vargas Koch
Verwandte Seminare
- Dagstuhl-Seminar 15412: Dynamic Traffic Models in Transportation Science (2015-10-04 - 2015-10-09) (Details)
- Dagstuhl-Seminar 18102: Dynamic Traffic Models in Transportation Science (2018-03-04 - 2018-03-09) (Details)
- Dagstuhl-Seminar 22192: Dynamic Traffic Models in Transportation Science (2022-05-08 - 2022-05-13) (Details)
- Dagstuhl-Seminar 24281: Dynamic Traffic Models in Transportation Science (2024-07-07 - 2024-07-12) (Details)
Klassifikation
- Computer Science and Game Theory
- Discrete Mathematics
- Multiagent Systems
Schlagworte
- dynamic equilibria
- dynamic traffic assignment
- traffic simulation
- equilibrium computation
- learning in games
