Research map / Multi-agent systems and games / Cooperative multi-agent RL
IMP-MARL: a Suite of Environments for Large-scale Infrastructure Management Planning via MARL
Pascal Leroy, Pablo G. Morato, Jonathan Pisane, Athanasios Kolios, Damien Ernst
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use MARL methods instead of expert heuristic policies | CTDE MARL methods · IMP-MARL environments | normalised discounted reward vs heuristic | Improved |
| use CTDE methods (QMIX, QVMIX, QPLEX) instead of independent Q-learning | IQL · IMP-MARL environments | normalised discounted reward vs heuristic | Improved |
| use CTDE methods instead of centralised DQN | DQN · IMP-MARL environments | normalised discounted reward vs heuristic | Improved |
| increase the number of agents for centralised DQN | DQN · IMP-MARL environments | normalised discounted reward vs heuristic | Hurt |
| add correlation between components in k-out-of-n systems | MARL methods · correlated k-out-of-n | normalised discounted reward vs heuristic | Hurt |
| add a global campaign cost with fewer than 50 agents | MARL methods · campaign cost environments | normalised discounted reward vs heuristic | Hurt |
| add a global campaign cost with more than 50 agents | MARL methods · campaign cost environments | normalised discounted reward vs heuristic | No clear effect |
| increase number of agents in offshore wind farm environments | MARL methods · offshore wind farm | normalised discounted reward vs heuristic | Hurt |
| increase the number of agents under campaign cost | MARL methods · campaign cost environments | normalised discounted reward vs heuristic | Improved |
| use COMA with more agents | COMA · IMP-MARL environments | normalised discounted reward vs heuristic | Hurt |
| use FACMAC with over 50 agents | FACMAC · IMP environments over 50 agents | normalised discounted reward vs heuristic | Improved |
Closest papers by meaning
- Benchmarking Multi-Agent Reinforcement Learning in Power Grid Operations (ICLR 2026 · similarity 0.88)
- Dispelling the Mirage of Progress in Offline MARL through Standardised Baselines and Evaluation (NeurIPS 2024 · similarity 0.85)
- Efficient Multi-agent Reinforcement Learning by Planning (ICLR 2024 · similarity 0.85)
- Exploring the Efficacy of Multi-Agent Reinforcement Learning for Autonomous Cyber Defence: A CAGE Challenge 4 Perspective (AAAI 2025 · similarity 0.84)
- Beyond Training Time, Test-Time Coordination is Essential for Cooperative MARL (NeurIPS 2026 · similarity 0.83)
- WFCRL: A Multi-Agent Reinforcement Learning Benchmark for Wind Farm Control (NeurIPS 2024 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
