Research map / Multi-agent systems and games / Cooperative multi-agent RL
MALib: A Parallel Framework for Population-based Multi-agent Reinforcement Learning
Ming Zhou, Ziyu Wan, Hanjing Wang, Muning Wen, Runzhe Wu, Ying Wen, Yaodong Yang, Yong Yu, Jun Wang, Weinan Zhang
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 alpha-rank as exact meta-solver in PSRO instead of fictitious play (ablation) | PSRO in MALib · Leduc Poker | time to exploitability 0.5 (sec): 4862 -> 1776 | Improved |
| use PSRO instead of Fictitious Self-Play | MALib population-based training · Leduc Poker | time to exploitability 0.5 (sec): 7562 -> 4862 | Improved |
| MALib implementation of PSRO instead of OpenSpiel | PSRO · Leduc Poker | execution time: cut 70% execution time while maintaining exploitability | Improved |
| MALib QMIX instead of PyMARL QMIX on 3m | QMIX · SMAC 3m | time to 80% win rate (sec): 5625 -> 300 | Improved |
| MALib QMIX instead of PyMARL QMIX on 8m | QMIX · SMAC 8m | time to 80% win rate (sec): 12375 -> 500 | Improved |
| MALib QMIX instead of PyMARL QMIX on 2s3z | QMIX · SMAC 2s3z | time to 80% win rate (sec): 7920 -> 375 | Improved |
| MALib asynchronous MADDPG instead of RLlib sequential MADDPG | MADDPG · MPE simple adversary | speedup: 5x speedup | Improved |
| MALib QMIX or PyMARL QMIX on 3s5z | QMIX · SMAC 3s5z | win rate | No clear effect |
Closest papers by meaning
- FightLadder: A Benchmark for Competitive Multi-Agent Reinforcement Learning (ICML 2024 · similarity 0.89)
- HyperMARL: Adaptive Hypernetworks for Multi-Agent RL (NeurIPS 2025 · similarity 0.88)
- Efficient Multi-agent Reinforcement Learning by Planning (ICLR 2024 · similarity 0.88)
- Dispelling the Mirage of Progress in Offline MARL through Standardised Baselines and Evaluation (NeurIPS 2024 · similarity 0.88)
- Foundations of Multi-Agent Learning in Dynamic Environments: Where Reinforcement Learning Meets Strategic Decision-Making (AAAI 2025 · similarity 0.87)
- JaxMARL: Multi-Agent RL Environments and Algorithms in JAX (NeurIPS 2024 · similarity 0.87)
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.
