IALight: Importance-Aware Multi-Agent Reinforcement Learning for Arterial Traffic Cooperative Control
Multi-intersection cooperative control for arterial or network scenarios is a crucial issue in urban traffic management. Multi-agent reinforcement learning (MARL) has been recognised as an efficient solution and shows outperformed results. However, most existing MARL-based methods treat intersection...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
University of Zagreb, Faculty of Transport and Traffic Sciences
2025-02-01
|
| Series: | Promet (Zagreb) |
| Subjects: | |
| Online Access: | https://traffic2.fpz.hr/index.php/PROMTT/article/view/650 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832086493865181184 |
|---|---|
| author | Lu WEI Xiaoyan ZHANG Lijun FAN Lei GAO Jian YANG |
| author_facet | Lu WEI Xiaoyan ZHANG Lijun FAN Lei GAO Jian YANG |
| author_sort | Lu WEI |
| collection | DOAJ |
| description | Multi-intersection cooperative control for arterial or network scenarios is a crucial issue in urban traffic management. Multi-agent reinforcement learning (MARL) has been recognised as an efficient solution and shows outperformed results. However, most existing MARL-based methods treat intersection equally, ignoring different importance of each intersection, such as high traffic volume, connecting multiple main roads, serving as entry or exit point for highways or commercial areas, etc. Besides, learning efficiency and practicality remain challenges. To address these issues, this paper proposes a novel importance-aware MARL-based method named IALight for traffic optimisation control. First, a normalised traffic pressure is introduced to ensure our state and reward design can accurately reflect the status of intersection traffic flow. Second, a reward adjustment module is designed to modify the reward based on intersection importance. To enhance practicality and safety for real-world applications, we adopt a green duration optimisation strategy under a cyclic fixed phase sequence. Comprehensive experiments on both synthetic and real-world traffic scenarios demonstrate that the proposed IALight outperforms the traditional and deep reinforcement learning baselines by more than 20.41% and 17.88% in average vehicle travel time, respectively. |
| format | Article |
| id | doaj-art-35f119d49ca44239a0cb558ff2642abf |
| institution | Kabale University |
| issn | 0353-5320 1848-4069 |
| language | English |
| publishDate | 2025-02-01 |
| publisher | University of Zagreb, Faculty of Transport and Traffic Sciences |
| record_format | Article |
| series | Promet (Zagreb) |
| spelling | doaj-art-35f119d49ca44239a0cb558ff2642abf2025-02-06T12:36:20ZengUniversity of Zagreb, Faculty of Transport and Traffic SciencesPromet (Zagreb)0353-53201848-40692025-02-0137115116910.7307/ptt.v37i1.650650IALight: Importance-Aware Multi-Agent Reinforcement Learning for Arterial Traffic Cooperative ControlLu WEI0Xiaoyan ZHANG1Lijun FAN2Lei GAO3Jian YANG4Beijing Polytechnic College, School of Information EngineeringBeijing Polytechnic College, School of Information EngineeringBeijing Polytechnic College, School of Information EngineeringNorth China University of Technology, School of Computer Science and TechnologyNorth China University of Technology, School of Computer Science and TechnologyMulti-intersection cooperative control for arterial or network scenarios is a crucial issue in urban traffic management. Multi-agent reinforcement learning (MARL) has been recognised as an efficient solution and shows outperformed results. However, most existing MARL-based methods treat intersection equally, ignoring different importance of each intersection, such as high traffic volume, connecting multiple main roads, serving as entry or exit point for highways or commercial areas, etc. Besides, learning efficiency and practicality remain challenges. To address these issues, this paper proposes a novel importance-aware MARL-based method named IALight for traffic optimisation control. First, a normalised traffic pressure is introduced to ensure our state and reward design can accurately reflect the status of intersection traffic flow. Second, a reward adjustment module is designed to modify the reward based on intersection importance. To enhance practicality and safety for real-world applications, we adopt a green duration optimisation strategy under a cyclic fixed phase sequence. Comprehensive experiments on both synthetic and real-world traffic scenarios demonstrate that the proposed IALight outperforms the traditional and deep reinforcement learning baselines by more than 20.41% and 17.88% in average vehicle travel time, respectively.https://traffic2.fpz.hr/index.php/PROMTT/article/view/650traffic signal controlintersection importancemulti-agent reinforcement learningarterial cooperative control |
| spellingShingle | Lu WEI Xiaoyan ZHANG Lijun FAN Lei GAO Jian YANG IALight: Importance-Aware Multi-Agent Reinforcement Learning for Arterial Traffic Cooperative Control Promet (Zagreb) traffic signal control intersection importance multi-agent reinforcement learning arterial cooperative control |
| title | IALight: Importance-Aware Multi-Agent Reinforcement Learning for Arterial Traffic Cooperative Control |
| title_full | IALight: Importance-Aware Multi-Agent Reinforcement Learning for Arterial Traffic Cooperative Control |
| title_fullStr | IALight: Importance-Aware Multi-Agent Reinforcement Learning for Arterial Traffic Cooperative Control |
| title_full_unstemmed | IALight: Importance-Aware Multi-Agent Reinforcement Learning for Arterial Traffic Cooperative Control |
| title_short | IALight: Importance-Aware Multi-Agent Reinforcement Learning for Arterial Traffic Cooperative Control |
| title_sort | ialight importance aware multi agent reinforcement learning for arterial traffic cooperative control |
| topic | traffic signal control intersection importance multi-agent reinforcement learning arterial cooperative control |
| url | https://traffic2.fpz.hr/index.php/PROMTT/article/view/650 |
| work_keys_str_mv | AT luwei ialightimportanceawaremultiagentreinforcementlearningforarterialtrafficcooperativecontrol AT xiaoyanzhang ialightimportanceawaremultiagentreinforcementlearningforarterialtrafficcooperativecontrol AT lijunfan ialightimportanceawaremultiagentreinforcementlearningforarterialtrafficcooperativecontrol AT leigao ialightimportanceawaremultiagentreinforcementlearningforarterialtrafficcooperativecontrol AT jianyang ialightimportanceawaremultiagentreinforcementlearningforarterialtrafficcooperativecontrol |