T-SCORE Project M2: Multi-Agent Simulation: A Model of Ride-Hailing Driver Participation

Tenenboim, Einat; Xu, Yufei; Macfarlane, Gregory S; Erhardt, Gregory D; Peeta, Srinivas · 2023 · ROSA P / Transit – Serving Communities Optimally, Responsively, and Efficiently (T-SCORE) Center [UTC]

archive: archived pipeline: cataloged verified

Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)

Summary

This study addresses a critical gap in ride-hailing research: the lack of understanding regarding driver supply behavior. While existing literature heavily focuses on trip demand, this project aims to model driver participation to enable realistic multi-agent simulations for urban transportation planning. The research was motivated by the need for transit agencies to understand how private ride-hailing services compete with or complement public transit, particularly given that driver decisions directly impact system performance and congestion. The researchers utilized a unique dataset of Uber and Lyft vehicle traces collected in San Francisco during November and December 2016. Due to the absence of unique driver identifiers in Uber data, the study relied on Lyft data for modeling driver participation, scaling results to account for Uber trips. The dataset included 15,670 Lyft drivers categorized by weekly working hours into occasional (<5 hours), part-time (5–35 hours), and full-time (>35 hours) groups. Using a discrete choice modeling approach, the team estimated driver participation across four sequential steps: (1) the number of shifts worked per day, (2) shift duration, (3) shift start time, and (4) shift start location. Shifts were defined as continuous working periods without breaks exceeding 60 minutes. Key findings indicate that driver type is a strong determinant of participation patterns. Full-time drivers were significantly more likely to work multiple shifts and longer durations compared to occasional drivers. Regarding timing, full-time drivers showed a uniform preference for starting shifts between 5 AM and 6 PM, whereas part-time drivers preferred early morning (2–9 AM) and late afternoon/evening (4–8 PM) starts. Location modeling revealed that shifts predominantly started in downtown areas with high population and employment density, as well as in higher-income neighborhoods. Conversely, fewer shifts started in areas with high college student density, likely due to students’ reliance on walking or micro-mobility. The models also confirmed that drivers tend to start subsequent shifts in districts bordering where their previous shift ended. The significance of this work lies in its contribution to the limited body of research on ride-hailing supply. By generating a realistic driver fleet for a typical weekday based on these behavioral models, the study provides a foundational tool for multi-agent simulations. These simulations can help transit agencies and policymakers better forecast ride-hailing operations, evaluate the impacts of ride-hailing on congestion and transit ridership, and design effective regulations for multi-modal transportation systems.

Key finding

Driver type is a strong determinant of ride-hailing participation, with full-time drivers working significantly more and longer shifts than occasional drivers, and shifts predominantly starting in high-population, high-employment, and high-income areas.

Methodology

modeling

Sample size: 15670

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• temporal

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Theoretical Contribution: computational model