A Robust and Efficient Multi-Agent Reinforcement Learning Framework for Traffic Signal Control

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Huang, Sheng-You, Chang, Hsiao-Chuan, Chen, Yen-Chi, Wei, Ting-Han, Yeh, I-Hau, Kuan, Sheng-Yao, Wang, Chien-Yao, Lee, Hsuan-Han, Wu, I-Chen
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915857685282816
author Huang, Sheng-You
Chang, Hsiao-Chuan
Chen, Yen-Chi
Wei, Ting-Han
Yeh, I-Hau
Kuan, Sheng-Yao
Wang, Chien-Yao
Lee, Hsuan-Han
Wu, I-Chen
author_facet Huang, Sheng-You
Chang, Hsiao-Chuan
Chen, Yen-Chi
Wei, Ting-Han
Yeh, I-Hau
Kuan, Sheng-Yao
Wang, Chien-Yao
Lee, Hsuan-Han
Wu, I-Chen
contents Reinforcement Learning (RL) in Traffic Signal Control (TSC) faces significant hurdles in real-world deployment due to limited generalization to dynamic traffic flow variations. Existing approaches often overfit static patterns and use action spaces incompatible with driver expectations. This paper proposes a robust Multi-Agent Reinforcement Learning (MARL) framework validated in the Vissim traffic simulator. The framework integrates three mechanisms: (1) Turning Ratio Randomization, a training strategy that exposes agents to dynamic turning probabilities to enhance robustness against unseen scenarios; (2) a stability-oriented Exponential Phase Duration Adjustment action space, which balances responsiveness and precision through cyclical, exponential phase adjustments; and (3) a Neighbor-Based Observation scheme utilizing the MAPPO algorithm with Centralized Training with Decentralized Execution (CTDE). By leveraging centralized updates, this approach approximates the efficacy of global observations while maintaining scalable local communication. Experimental results demonstrate that our framework outperforms standard RL baselines, reducing average waiting time by over 10%. The proposed model exhibits superior generalization in unseen traffic scenarios and maintains high control stability, offering a practical solution for adaptive signal control.
format Preprint
id arxiv_https___arxiv_org_abs_2603_12096
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle A Robust and Efficient Multi-Agent Reinforcement Learning Framework for Traffic Signal Control
Huang, Sheng-You
Chang, Hsiao-Chuan
Chen, Yen-Chi
Wei, Ting-Han
Yeh, I-Hau
Kuan, Sheng-Yao
Wang, Chien-Yao
Lee, Hsuan-Han
Wu, I-Chen
Artificial Intelligence
Reinforcement Learning (RL) in Traffic Signal Control (TSC) faces significant hurdles in real-world deployment due to limited generalization to dynamic traffic flow variations. Existing approaches often overfit static patterns and use action spaces incompatible with driver expectations. This paper proposes a robust Multi-Agent Reinforcement Learning (MARL) framework validated in the Vissim traffic simulator. The framework integrates three mechanisms: (1) Turning Ratio Randomization, a training strategy that exposes agents to dynamic turning probabilities to enhance robustness against unseen scenarios; (2) a stability-oriented Exponential Phase Duration Adjustment action space, which balances responsiveness and precision through cyclical, exponential phase adjustments; and (3) a Neighbor-Based Observation scheme utilizing the MAPPO algorithm with Centralized Training with Decentralized Execution (CTDE). By leveraging centralized updates, this approach approximates the efficacy of global observations while maintaining scalable local communication. Experimental results demonstrate that our framework outperforms standard RL baselines, reducing average waiting time by over 10%. The proposed model exhibits superior generalization in unseen traffic scenarios and maintains high control stability, offering a practical solution for adaptive signal control.
title A Robust and Efficient Multi-Agent Reinforcement Learning Framework for Traffic Signal Control
topic Artificial Intelligence
url https://arxiv.org/abs/2603.12096