Thesis Title: Data-Driven Network Design of On-Demand Multimodal Transit Systems

Thesis Committee:

Dr. Pascal Van Hentenryck (advisor), Industrial and Systems Engineering, Georgia Institute of Technology

Dr. Alan Erera, Industrial and Systems Engineering, Georgia Institute of Technology

Dr. Alejandro Toriello, Industrial and Systems Engineering, Georgia Institute of Technology

Dr. Chelsea (Chip) White, Industrial and Systems Engineering, Georgia Institute of Technology

Dr. Kari Watkins, Civil and Environmental Engineering, University of California Davis

Date and Time: Monday, December 5th, 2pm (EST)

In-Person Location: Midtown Room (Coda C12-15)

Meeting Link: https://teams.microsoft.com/l/meetup-join/19%3ameeting_OGUwNDMyOGMtODdhNy00NzhmLTgyOTctMWZlNjFhYmZkODU4%40thread.v2/0?context=%7b%22Tid%22%3a%22482198bb-ae7b-4b25-8b7a-6d7f32faa083%22%2c%22Oid%22%3a%22ff916efd-6d1b-468f-a9cc-d7bb53e55642%22%7d

Meeting ID: 275 429 627 179
Passcode: 4skjpY

Abstract:

Across the United States, public transit agencies are facing trends of decreasing ridership. Especially during and since the COVID-19 pandemic, reduced ridership caused many agencies to have significant budget deficits due the to the high-fixed cost of traditional transit systems. Many transit planners are exploring network redesign to address these changing ridership patterns and their budget deficits. On-demand services have previous been used in smaller cities, and more rural areas and for paratransit services, but now larger cities are starting to explore using on-demand services to help supplement their fixed route services to create a more accessible and scalable system. On-demand Multimodal Transit Systems may be an accessible, scalable solution for first and last mile issues that often plague many transportation systems.

Chapter 2 presents a novel methodology to help transit agencies with their tactical planning for post-game ridership of large events. The methodology has three main steps: (1) predicting the total post-game ridership; (2) combining the total prediction with historical trends to forecast the passenger flow curve at nearby stations after the game; and (3) estimating the required train frequencies to serve these customers with minimal passengers left behind by each train. Additionally, this chapter proposes a suite of data-driven techniques that together create a data-driven pipeline to exploit Automated Fare Collection (AFC) data for evaluating, anticipating, and managing the performance of

transit systems. This chapter includes a case study where the proposed pipeline is used to generate an adjusted train schedule for the post-game period and simulated with the rail ridership data from the Metropolitan Atlanta Rapid Transit Authority (MARTA). The simulation results highlight how the proposed schedules based on the estimated required post-game train frequencies could significantly improve post-game congestion and wait time.

Chapter 3 studies the resiliency during a pandemic of On-Demand Multimodal Transit Systems (ODMTS), a new generation of transit systems that combine a network of high-frequency trains and buses with on-demand shuttles to serve the first and last miles and act as feeders to the fixed network. It presents a case study for the city of Atlanta and evaluates ODMTS for multiple scenarios of depressed demand and social distancing representing various stages of the pandemic. The case study relies on a real data from MARTA, an optimization pipeline for the design of ODMTS, and a detailed simulation of these designs. The case study demonstrates how ODMTS provide a resilient solution in terms of cost, convenience, and accessibility for this wide range of scenarios.

Chapter 4 studies to what extent these benefits apply to On-Demand Multimodal Transit Systems (ODMTS); a novel type of transit systems that combines traditional rail and bus networks with on-demand shuttles. Previous case studies have shown that ODMTS may simultaneously improve travel time, reduce system cost, and attract new passengers compared to the existing systems. However, none of these studies include the effect of congestion on the performance of the ODMTS. This chapter models the ODMTS with multiple congestion scenarios as part of a case study in the Metro Atlanta Area, both with and without DBLs. The overall assessment of the case study reveals that DBLs in combination with an ODMTS could attract even more ridership adoption by mitigating the negative impact of congestion on the ODMTS with DBLs on the most congested corridor.

Chapter 5 presents a novel mixed-integer program (MIP) formulation to incorporate bus line design into the network design problem for On-Demand Multimodal Transit Systems (ODMTS) that allow the model to accurately capture wait time and transfer costs in addition to travel time and vehicle costs. To solve large-scale instances, a two-stage reformulation is presented where the first-stage problem decides which bus arcs to open and decides which arcs immediately follow and the second-stage problem decides the multimodal path for each individual trip. The solution method is based on the Benders decomposition method and uses disaggregated subproblems and Pareto-optimal cuts. This chapter includes a case study of the Metro Atlanta Area with instances that have up to 43,000 unique trips and hundreds of bus arcs for potential lines. The results show there was a significant reduction in number of transfers for individuals when considering the bus lines as part of the network design phase.