Model-Based Graph Reinforcement Learning for Inductive Traffic Signal Control
We introduce MuJAM, an adaptive traffic signal control method which leverages model-based reinforcement learning to 1) extend recent generalization efforts (to road network architectures and traffic distributions) further by allowing a generalization to the controllers’ constraints (cycli...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Open Journal of Intelligent Transportation Systems |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10470423/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832590344096579584 |
|---|---|
| author | Francois-Xavier Devailly Denis Larocque Laurent Charlin |
| author_facet | Francois-Xavier Devailly Denis Larocque Laurent Charlin |
| author_sort | Francois-Xavier Devailly |
| collection | DOAJ |
| description | We introduce MuJAM, an adaptive traffic signal control method which leverages model-based reinforcement learning to 1) extend recent generalization efforts (to road network architectures and traffic distributions) further by allowing a generalization to the controllers’ constraints (cyclic and acyclic policies), 2) improve performance and data efficiency over related model-free approaches, and 3) enable explicit coordination at scale for the first time. In a zero-shot transfer setting involving both road networks and traffic settings never experienced during training, and in a larger transfer experiment involving the control of 3,971 traffic signal controllers in Manhattan, we show that MuJAM, using both cyclic and acyclic constraints, outperforms domain-specific baselines as well as a recent transferable approach. |
| format | Article |
| id | doaj-art-92d2952a76ca4d67b5b5d94108b55636 |
| institution | Kabale University |
| issn | 2687-7813 |
| language | English |
| publishDate | 2024-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Open Journal of Intelligent Transportation Systems |
| spelling | doaj-art-92d2952a76ca4d67b5b5d94108b556362025-01-24T00:02:37ZengIEEEIEEE Open Journal of Intelligent Transportation Systems2687-78132024-01-01523825010.1109/OJITS.2024.337658310470423Model-Based Graph Reinforcement Learning for Inductive Traffic Signal ControlFrancois-Xavier Devailly0https://orcid.org/0000-0002-5861-0675Denis Larocque1https://orcid.org/0000-0002-7372-7943Laurent Charlin2https://orcid.org/0000-0002-6545-9459Department of Decision Sciences, HEC Montreal, Montreal, QC, CanadaDepartment of Decision Sciences, HEC Montreal, Montreal, QC, CanadaDepartment of Decision Sciences, HEC Montreal, Montreal, QC, CanadaWe introduce MuJAM, an adaptive traffic signal control method which leverages model-based reinforcement learning to 1) extend recent generalization efforts (to road network architectures and traffic distributions) further by allowing a generalization to the controllers’ constraints (cyclic and acyclic policies), 2) improve performance and data efficiency over related model-free approaches, and 3) enable explicit coordination at scale for the first time. In a zero-shot transfer setting involving both road networks and traffic settings never experienced during training, and in a larger transfer experiment involving the control of 3,971 traffic signal controllers in Manhattan, we show that MuJAM, using both cyclic and acyclic constraints, outperforms domain-specific baselines as well as a recent transferable approach.https://ieeexplore.ieee.org/document/10470423/Adaptive traffic signal controltransfer learningmulti-agent reinforcement learningjoint action modelingmodel-based reinforcement learninggraph neural networks |
| spellingShingle | Francois-Xavier Devailly Denis Larocque Laurent Charlin Model-Based Graph Reinforcement Learning for Inductive Traffic Signal Control IEEE Open Journal of Intelligent Transportation Systems Adaptive traffic signal control transfer learning multi-agent reinforcement learning joint action modeling model-based reinforcement learning graph neural networks |
| title | Model-Based Graph Reinforcement Learning for Inductive Traffic Signal Control |
| title_full | Model-Based Graph Reinforcement Learning for Inductive Traffic Signal Control |
| title_fullStr | Model-Based Graph Reinforcement Learning for Inductive Traffic Signal Control |
| title_full_unstemmed | Model-Based Graph Reinforcement Learning for Inductive Traffic Signal Control |
| title_short | Model-Based Graph Reinforcement Learning for Inductive Traffic Signal Control |
| title_sort | model based graph reinforcement learning for inductive traffic signal control |
| topic | Adaptive traffic signal control transfer learning multi-agent reinforcement learning joint action modeling model-based reinforcement learning graph neural networks |
| url | https://ieeexplore.ieee.org/document/10470423/ |
| work_keys_str_mv | AT francoisxavierdevailly modelbasedgraphreinforcementlearningforinductivetrafficsignalcontrol AT denislarocque modelbasedgraphreinforcementlearningforinductivetrafficsignalcontrol AT laurentcharlin modelbasedgraphreinforcementlearningforinductivetrafficsignalcontrol |