Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
July 15, 2026Transportmetrica A Transport Science

Dynamic control of schedule-based and headway-based bus services using deep reinforcement learning

View Full Paper
Ask AI
Bookmark
Share

Authors

SAShervin AtaeianSSSaeid Saidi

Discussion

Loading...

Member takes

Overview

Randomized trial evaluates deep reinforcement learning for optimizing bus control in varying traffic conditions, suggesting improved efficiency.

Key Points

  • This study aims to enhance bus service reliability and efficiency using deep reinforcement learning techniques.
  • Utilized a Double Deep Q-Network architecture for bus control optimization.
  • Compared deep reinforcement learning agents against a no-control and rule-based strategy across schedule-based and headway-based services.
  • Evaluated control actions such as holding, stop-skipping, and speed adjustment under regular and severely disrupted conditions.
  • In regular operations, DRL integrated control reduced total passenger journey time median by up to 34%.
  • Under severe disruptions, DRL agents achieved a 24% median journey time reduction using stop-skipping in low-frequency scenarios.
  • Integrated control prevented bus bunching in high-frequency scenarios, balancing passenger loads.

Cite This Study

Ataeian et al. (2026) studied this question.

synapsesocial.com/papers/6a57231a88b21df87547ff0ehttps://doi.org/10.1080/23249935.2026.2699198
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Single Agent Robust Deep Reinforcement Learning for Bus Fleet Control2025
  2. 2Single Agent Robust Deep Reinforcement Learning for Bus Fleet Control2026
  3. 3Adaptive bus service scheduling with stochastic running times via an adversarial reinforcement learning approach2026
  4. 4Joint control of traffic signal phase sequence and timing: a deep reinforcement learning method2025 · 8 citations
  5. 5Dynamic Traffic Management Through Deep Reinforcement Learning2026