Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Forecasting and Mitigating Disruptions in Public Bus Transit Services

Chaeeun Han, Jose Paolo Talusan, Dan Freudberg, Ayan Mukhopadhyay, Abhishek Dubey, Aron Laszka

TL;DR

This paper tackles proactive disruption management in a mid-sized city by jointly forecasting bus-trip disruptions and optimally stationing substitute buses. It combines data-driven disruption probability models (logistic regression and XGBoost, with isotonic calibration) with a greedy + simulated-annealing optimization to select stationing stops, guided by a stochastic simulator that evaluates $D$ (non-service miles), $T$ (non-service time), and $L$ (passengers left behind). The approach is validated on Nashville data (GTFS, APC, disruption logs, weather) and through an event-driven simulator that generates 100 event chains per day across multiple days, showing reductions in left-behind passengers while maintaining deadhead metrics. The results demonstrate that forecast-informed stationing plans can outperform ad-hoc strategies, enabling more reliable and equitable public transit with scalable computation suitable for real-world decision support.

Abstract

Public transportation systems often suffer from unexpected fluctuations in demand and disruptions, such as mechanical failures and medical emergencies. These fluctuations and disruptions lead to delays and overcrowding, which are detrimental to the passengers' experience and to the overall performance of the transit service. To proactively mitigate such events, many transit agencies station substitute (reserve) vehicles throughout their service areas, which they can dispatch to augment or replace vehicles on routes that suffer overcrowding or disruption. However, determining the optimal locations where substitute vehicles should be stationed is a challenging problem due to the inherent randomness of disruptions and due to the combinatorial nature of selecting locations across a city. In collaboration with the transit agency of Nashville, TN, we address this problem by introducing data-driven statistical and machine-learning models for forecasting disruptions and an effective randomized local-search algorithm for selecting locations where substitute vehicles are to be stationed. Our research demonstrates promising results in proactive disruption management, offering a practical and easily implementable solution for transit agencies to enhance the reliability of their services. Our results resonate beyond mere operational efficiency: by advancing proactive strategies, our approach fosters more resilient and accessible public transportation, contributing to equitable urban mobility and ultimately benefiting the communities that rely on public transportation the most.

Forecasting and Mitigating Disruptions in Public Bus Transit Services

TL;DR

This paper tackles proactive disruption management in a mid-sized city by jointly forecasting bus-trip disruptions and optimally stationing substitute buses. It combines data-driven disruption probability models (logistic regression and XGBoost, with isotonic calibration) with a greedy + simulated-annealing optimization to select stationing stops, guided by a stochastic simulator that evaluates (non-service miles), (non-service time), and (passengers left behind). The approach is validated on Nashville data (GTFS, APC, disruption logs, weather) and through an event-driven simulator that generates 100 event chains per day across multiple days, showing reductions in left-behind passengers while maintaining deadhead metrics. The results demonstrate that forecast-informed stationing plans can outperform ad-hoc strategies, enabling more reliable and equitable public transit with scalable computation suitable for real-world decision support.

Abstract

Public transportation systems often suffer from unexpected fluctuations in demand and disruptions, such as mechanical failures and medical emergencies. These fluctuations and disruptions lead to delays and overcrowding, which are detrimental to the passengers' experience and to the overall performance of the transit service. To proactively mitigate such events, many transit agencies station substitute (reserve) vehicles throughout their service areas, which they can dispatch to augment or replace vehicles on routes that suffer overcrowding or disruption. However, determining the optimal locations where substitute vehicles should be stationed is a challenging problem due to the inherent randomness of disruptions and due to the combinatorial nature of selecting locations across a city. In collaboration with the transit agency of Nashville, TN, we address this problem by introducing data-driven statistical and machine-learning models for forecasting disruptions and an effective randomized local-search algorithm for selecting locations where substitute vehicles are to be stationed. Our research demonstrates promising results in proactive disruption management, offering a practical and easily implementable solution for transit agencies to enhance the reliability of their services. Our results resonate beyond mere operational efficiency: by advancing proactive strategies, our approach fosters more resilient and accessible public transportation, contributing to equitable urban mobility and ultimately benefiting the communities that rely on public transportation the most.
Paper Structure (21 sections, 3 equations, 8 figures, 6 tables, 3 algorithms)

This paper contains 21 sections, 3 equations, 8 figures, 6 tables, 3 algorithms.

Table of Contents

  1. Introduction
  2. Problem Formulation
  3. Disruption Forecasting
  4. Vehicle Stationing
  5. Proposed Approach
  6. Disruption Forecasting
  7. Vehicle Stationing
  8. Experimental Setup
  9. Real-World Data
  10. Disruption Forecasting
  11. Simulator
  12. Numerical Results
  13. Disruption Forecasting
  14. Model Performance
  15. Model and Feature Selection
  16. ...and 6 more sections

Figures (8)

  • Figure 1: Disruptions, such as traffic accidents, hamper the reliability of the service provided by our partner transit agency. This image shows an actual disruption from 2020.
  • Figure 2: Log-Odds and Probability Values for Features.
  • Figure 3: Feature importance of one-hot encoded categorical features used for XGBoost model.
  • Figure 4: Comparison of cost using stationing plans based on search, garage, or agency. Lines and shaded regions represent mean and standard error across 100 chains.
  • Figure 5: Normalized objective costs for all test days.
  • ...and 3 more figures