Design of Transit-Centric Multimodal Urban Mobility System with Autonomous Mobility-on-Demand

Xiaotong Guo, Jinhua ZhaoarXiv 2024 · OpenReview

Abstract

This paper addresses the pressing challenge of urban mobility in the context of growing urban populations, changing demand patterns for urban mobility, and emerging technologies like Mobility-on-Demand (MoD) platforms and Autonomous Vehicle (AV). As urban areas swell and demand pattern changes, the integration of Autonomous Mobility-on-Demand (AMoD) systems with existing public transit (PT) networks presents great opportunities to enhancing urban mobility. We propose a novel optimization framework for solving the Transit-Centric Multimodal Urban Mobility with Autonomous Mobility-on-Demand (TCMUM-AMoD) at scale. The system operator (public transit agency) determines the network design and frequency settings of the PT network, fleet sizing and allocations of AMoD system, and the pricing for using the multimodal system with the goal of minimizing passenger disutility. Passengers' mode and route choice behaviors are modeled explicitly using discrete choice models. A first-order approximation algorithm is introduced to solve the problem at scale. Using a case study in Chicago, we showcase the potential to optimize urban mobility across different demand scenarios. To our knowledge, ours is the first paper to jointly optimize transit network design, fleet sizing, and pricing for the multimodal mobility system while considering passengers' mode and route choices.

Introduction

Urban mobility is at a crossroads. As cities grow and post-pandemic travel patterns shift toward remote work, traditional public transit (PT) often struggles with the "first-mile-last-mile" (FMLM) problem—the difficulty of getting passengers from their front door to a high-capacity rail station. While Mobility-on-Demand (MoD) services like Uber and Lyft offer flexibility, they are often expensive and can increase congestion. Emerging Autonomous Vehicle (AV) technology offers a potential solution, but most research treats AVs and buses as rivals.

This paper addresses the challenge of joint optimization: How can a transit agency design a system where rail, buses, and autonomous fleets work together seamlessly? As shown in the literature comparison below, this study is unique in its joint consideration of transit network design, fleet sizing, and pricing while explicitly modeling passenger choice.

Comparison of this study with existing literature on multimodal transit and MoD optimization.

Methodology

The authors propose the Transit-Centric Multimodal Urban Mobility with Autonomous Mobility-on-Demand (TCMUM-AMoD) framework. The core of their approach is a Mixed Integer Non-Linear Program (MINLP) that balances three main levers: transit design (frequencies), AMoD management (fleet sizing and allocation), and pricing (discount factors).

The framework considers various multimodal paths to address the FMLM problem, integrating AMoD as a feeder to high-capacity rail.

Multimodal travel options including rail-only, bus-rail, and AMoD-rail combinations.

To make the model realistic, they incorporate Discrete Choice Models (specifically Multinomial and Nested Logit models) to predict how passengers will actually behave. Because these models introduce complex non-linearities that make the problem intractable at city-scale, the authors developed a first-order approximation algorithm. This algorithm replaces the non-linear functions with a series of local linear approximations, allowing the system to be solved iteratively and efficiently. The framework was validated using a case study of the Chicago Transit Authority (CTA) network, specifically focusing on the southern region of Chicago and the Red Line, utilizing GTFS, AVL, and ODX datasets.

Results and Discussion

The study reveals a "sweet spot" for integration: AMoD is most effective when it replaces infrequent, low-demand bus routes and acts as a high-frequency feeder to rail stations. However, for high-density "downtown" demand, traditional buses remain more efficient due to their high capacity.

Key findings from the Chicago case study include:

  • Efficiency Gains: In a Capital Cost Equivalence (CCE) scenario, replacing 20% of bus runs with 162 AMoD vehicles resulted in a 15.8% reduction in total passenger disutility.
  • Pricing Pitfalls: Surprisingly, the model suggests that discounting AMoD fares might be counterproductive. If AMoD becomes too attractive, it can lead to an "upsurge in excess waiting time" as the limited fleet becomes overwhelmed, actually increasing total system disutility.
  • Service Balance: Adding more AMoD vehicles doesn't always lead to a linear decrease in bus usage. As AMoD wait times drop, more people switch to it, which can necessitate a baseline level of bus service to maintain balance for local commuters.

Conclusion

While the TCMUM-AMoD framework offers a robust blueprint for the next generation of urban transit, the authors acknowledge several limitations. Current models assume static demand and fixed travel times, omitting real-time traffic dynamics and vehicle rebalancing ("deadheading"), which are major operational costs. Future research aims to incorporate these dynamic factors to create an even more robust blueprint for multimodal urban mobility.

Summary updated

Abstract

This paper addresses the pressing challenge of urban mobility in the context of growing urban populations, changing demand patterns for urban mobility, and emerging technologies like Mobility-on-Demand (MoD) platforms and Autonomous Vehicle (AV). As urban areas swell and demand pattern changes, the integration of Autonomous Mobility-on-Demand (AMoD) systems with existing public transit (PT) networks presents great opportunities to enhancing urban mobility. We propose a novel optimization framework for solving the Transit-Centric Multimodal Urban Mobility with Autonomous Mobility-on-Demand (TCMUM-AMoD) at scale. The system operator (public transit agency) determines the network design and frequency settings of the PT network, fleet sizing and allocations of AMoD system, and the pricing for using the multimodal system with the goal of minimizing passenger disutility. Passengers' mode and route choice behaviors are modeled explicitly using discrete choice models. A first-order approximation algorithm is introduced to solve the problem at scale. Using a case study in Chicago, we showcase the potential to optimize urban mobility across different demand scenarios. To our knowledge, ours is the first paper to jointly optimize transit network design, fleet sizing, and pricing for the multimodal mobility system while considering passengers' mode and route choices.
BibTeX
@misc{guo2024design,
  author = {Xiaotong Guo and Jinhua Zhao},
  title = {{Design of Transit-Centric Multimodal Urban Mobility System with Autonomous Mobility-on-Demand}},
  year = {2024},
  eprint = {2404.05885},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2404.05885},
}
Figure