Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Estimating and Mitigating the Congestion Effect of Curbside Pick-ups and Drop-offs: A Causal Inference Approach

Xiaohui Liu, Sean Qian, Hock-Hai Teo, Wei Ma

TL;DR

This study tackles curbside congestion from PUDOs by developing a causal-inference framework that links NoPUDO to traffic speed via a spatio-temporal causal graph. It introduces DSML, which decouples speed and PUDO dynamics into three sub-models to unbiasedly estimate region-specific congestion effects $\theta_v$, with data-splitting ensuring causal identification. A re-routing optimization leverages these estimates to redistribute PUDOs to nearby regions, achieving measurable reductions in network-wide total travel time in Manhattan (e.g., weekday Midtown $\Delta\mathrm{TTT}$ about $-2.44\%$, Central Park $-2.12\%$). The empirical results, robustness checks, and sensitivity analyses demonstrate practical applicability and resilience of the approach, offering a principled method for curbside management and policy design.

Abstract

Curb space is one of the busiest areas in urban road networks. Especially in recent years, the rapid increase of ride-hailing trips and commercial deliveries has induced massive pick-ups/drop-offs (PUDOs), which occupy the limited curb space that was designed and built decades ago. These PUDOs could jam curbside utilization and disturb the mainline traffic flow, evidently leading to significant negative societal externalities. However, there is a lack of an analytical framework that rigorously quantifies and mitigates the congestion effect of PUDOs in the system view, particularly with little data support and involvement of confounding effects. To bridge this research gap, this paper develops a rigorous causal inference approach to estimate the congestion effect of PUDOs on general regional networks. A causal graph is set to represent the spatio-temporal relationship between PUDOs and traffic speed, and a double and separated machine learning (DSML) method is proposed to quantify how PUDOs affect traffic congestion. Additionally, a re-routing formulation is developed and solved to encourage passenger walking and traffic flow re-routing to achieve system optimization. Numerical experiments are conducted using real-world data in the Manhattan area. On average, 100 additional units of PUDOs in a region could reduce the traffic speed by 3.70 and 4.54 mph on weekdays and weekends, respectively. Re-routing trips with PUDOs on curb space could respectively reduce the system-wide total travel time by 2.44% and 2.12% in Midtown and Central Park on weekdays. Sensitivity analysis is also conducted to demonstrate the effectiveness and robustness of the proposed framework.

Estimating and Mitigating the Congestion Effect of Curbside Pick-ups and Drop-offs: A Causal Inference Approach

TL;DR

This study tackles curbside congestion from PUDOs by developing a causal-inference framework that links NoPUDO to traffic speed via a spatio-temporal causal graph. It introduces DSML, which decouples speed and PUDO dynamics into three sub-models to unbiasedly estimate region-specific congestion effects , with data-splitting ensuring causal identification. A re-routing optimization leverages these estimates to redistribute PUDOs to nearby regions, achieving measurable reductions in network-wide total travel time in Manhattan (e.g., weekday Midtown about , Central Park ). The empirical results, robustness checks, and sensitivity analyses demonstrate practical applicability and resilience of the approach, offering a principled method for curbside management and policy design.

Abstract

Curb space is one of the busiest areas in urban road networks. Especially in recent years, the rapid increase of ride-hailing trips and commercial deliveries has induced massive pick-ups/drop-offs (PUDOs), which occupy the limited curb space that was designed and built decades ago. These PUDOs could jam curbside utilization and disturb the mainline traffic flow, evidently leading to significant negative societal externalities. However, there is a lack of an analytical framework that rigorously quantifies and mitigates the congestion effect of PUDOs in the system view, particularly with little data support and involvement of confounding effects. To bridge this research gap, this paper develops a rigorous causal inference approach to estimate the congestion effect of PUDOs on general regional networks. A causal graph is set to represent the spatio-temporal relationship between PUDOs and traffic speed, and a double and separated machine learning (DSML) method is proposed to quantify how PUDOs affect traffic congestion. Additionally, a re-routing formulation is developed and solved to encourage passenger walking and traffic flow re-routing to achieve system optimization. Numerical experiments are conducted using real-world data in the Manhattan area. On average, 100 additional units of PUDOs in a region could reduce the traffic speed by 3.70 and 4.54 mph on weekdays and weekends, respectively. Re-routing trips with PUDOs on curb space could respectively reduce the system-wide total travel time by 2.44% and 2.12% in Midtown and Central Park on weekdays. Sensitivity analysis is also conducted to demonstrate the effectiveness and robustness of the proposed framework.
Paper Structure (38 sections, 5 theorems, 33 equations, 18 figures, 5 tables, 2 algorithms)

This paper contains 38 sections, 5 theorems, 33 equations, 18 figures, 5 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Motivation
  3. Challenges and opportunities
  4. Contributions
  5. Model
  6. Causal relationship between NoPUDO and Traffic Speed
  7. Structural causal modeling
  8. Structural equation models
  9. Double and Separated Machine Learning
  10. Model $\mathtt{Y}$
  11. Model $\mathtt{D}$
  12. Model $\mathtt{Z}$
  13. Re-routing traffic flow with PUDOs to reduce total travel time
  14. Solution algorithms
  15. Solving the DSML method
  16. ...and 23 more sections

Key Result

Proposition 1

Suppose that Equation eq:y, eq:d, eq:noise1, and eq:noise2 hold and $y_v^t$, $d_v^t$, and $\mathbf{W}_v^t$ are observable for all $v, t$, then $\theta_v$ is identifiable, i.e., $\theta_v$ can be uniquely estimated from $y_v^t, d_v^t, \mathbf{W}_v^t, \forall v, t$.

Figures (18)

  • Figure 1: Illustration of congestion effect caused by PUDOs.
  • Figure 2: Relationship among travel demands, NoPUDO and traffic speed.
  • Figure 3: The causal graph of the NoPUDO and traffic speed.
  • Figure 4: The framework of the DSML method.
  • Figure 5: An example of the traffic flow re-routing with PUDOs.
  • ...and 13 more figures

Theorems & Definitions (10)

  • Example 1
  • Remark 1: Constant effects within a region
  • Remark 2: Heterogeneous effects across different regions
  • Remark 3: Independent effects across different regions
  • Proposition 1: Identifiable
  • Proposition 2
  • Proposition 3: FWL Theorem
  • Proposition 4
  • Example 2
  • Proposition 5: Total travel time decomposition