Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Dynamic Replanning for Improved Public Transport Routing

Abdallah Abuaisha, Bojie Shen, Daniel Harabor, Peter Stuckey, Mark Wallace

TL;DR

The paper addresses delays in public transport and proposes dynamic replanning as a scalable solution. It formalizes the problem and introduces two architectures: a Pull approach (server-driven replanning on demand) and a Push approach (proactive, envelope-restricted replanning executed with edge devices). The Push method leverages a Time-Independent Graph (TIG) and the envelope concept to bound search, enabling CSA-based replanning to run in sub-second times and handle dramatically more queries. Empirical results across four GTFS-based metropolitan networks show that dynamic replanning reduces travel times by 5–30 minutes on average for affected journeys, with the Push approach delivering substantial speedups and higher scalability. The work demonstrates the practicality of system-scale dynamic replanning for real-time public transport routing and paves the way for robust, delay-aware journey optimization in urban networks.

Abstract

Delays in public transport are common, often impacting users through prolonged travel times and missed transfers. Existing solutions for handling delays remain limited; backup plans based on historical data miss opportunities for earlier arrivals, while snapshot planning accounts for current delays but not future ones. With the growing availability of live delay data, users can adjust their journeys in real-time. However, the literature lacks a framework that fully exploits this advantage for system-scale dynamic replanning. To address this, we formalise the dynamic replanning problem in public transport routing and propose two solutions: a "pull" approach, where users manually request replanning, and a novel "push" approach, where the server proactively monitors and adjusts journeys. Our experiments show that the push approach outperforms the pull approach, achieving significant speedups. The results also reveal substantial arrival time savings enabled by dynamic replanning.

Dynamic Replanning for Improved Public Transport Routing

TL;DR

The paper addresses delays in public transport and proposes dynamic replanning as a scalable solution. It formalizes the problem and introduces two architectures: a Pull approach (server-driven replanning on demand) and a Push approach (proactive, envelope-restricted replanning executed with edge devices). The Push method leverages a Time-Independent Graph (TIG) and the envelope concept to bound search, enabling CSA-based replanning to run in sub-second times and handle dramatically more queries. Empirical results across four GTFS-based metropolitan networks show that dynamic replanning reduces travel times by 5–30 minutes on average for affected journeys, with the Push approach delivering substantial speedups and higher scalability. The work demonstrates the practicality of system-scale dynamic replanning for real-time public transport routing and paves the way for robust, delay-aware journey optimization in urban networks.

Abstract

Delays in public transport are common, often impacting users through prolonged travel times and missed transfers. Existing solutions for handling delays remain limited; backup plans based on historical data miss opportunities for earlier arrivals, while snapshot planning accounts for current delays but not future ones. With the growing availability of live delay data, users can adjust their journeys in real-time. However, the literature lacks a framework that fully exploits this advantage for system-scale dynamic replanning. To address this, we formalise the dynamic replanning problem in public transport routing and propose two solutions: a "pull" approach, where users manually request replanning, and a novel "push" approach, where the server proactively monitors and adjusts journeys. Our experiments show that the push approach outperforms the pull approach, achieving significant speedups. The results also reveal substantial arrival time savings enabled by dynamic replanning.

Paper Structure

This paper contains 23 sections, 1 theorem, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Preliminaries
  3. Timetable Modelling
  4. Earliest Arrival Time Problem (EATP).
  5. Delay Modelling
  6. Snapshot Earliest Arrival Time Problem (SEATP).
  7. Snapshot Solver
  8. Dynamic Replanning
  9. Our Approaches
  10. The Pull Approach
  11. The Push Approach
  12. Building the Envelope
  13. Replanning on the Envelope
  14. Experiments
  15. Experiments Setup
  16. ...and 8 more sections

Key Result

Theorem 1

There is no journey $j(s_o, s_d)$ from $s_o$ to $s_d$, starting at or after $\tau_q$ and using connections outside the envelope $Env$, that can arrive before $\tau_d$, regardless of delays.

Figures (4)

  • Figure 1: A toy network with eight stops ($s_1$ to $s_8$), where the origin $s_1$ is yellow and the destination $s_6$ is pink. Routes $r_1$, $r_2$, and $r_3$ use dashed blue, solid red, and dotted green arrows, labelled with travel times in minutes. Scheduled departure times for trips on each route are noted at its starting stop. Loops at stops indicate transfer times.
  • Figure 2: Framework for dynamic replanning during a typical replan step: pull approach (left) and push approach (right).
  • Figure 3: Runtime comparison between the pull and push approaches throughout the day. Different y-axis limits are used to enhance data clarity across plots. Cyan blocks represent peak periods, while grey lines indicate overall average runtimes.
  • Figure 4: Distribution of difference in arrival time at destination for affected queries (percentages shown) between static planning (SP) and dynamic replanning (DR) across the day. Red dots indicate mean values, while cyan blocks indicate peak periods.

Theorems & Definitions (8)

  • Example 1
  • Definition 1
  • Definition 2
  • Example 2
  • Definition 3
  • Example 3
  • Theorem 1
  • proof