Research map / Multi-agent systems and games / Cooperative multi-agent RL
Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning
Tabish Rashid, Mikayel Samvelyan, Christian Schroeder de Witt, Gregory Farquhar, Jakob Foerster, Shimon Whiteson
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 |
|---|---|---|---|
| add monotonic state-conditioned mixing network (QMIX vs VDN) | VDN · SMAC 3s5z | test win rate (%): 84 -> 97 | Improved |
| add monotonic mixing network (QMIX vs VDN) | VDN · SMAC 2c_vs_64zg | test win rate (%): 21 -> 58 | Improved |
| add monotonic mixing network (QMIX vs VDN) | VDN · SMAC MMM2 | test win rate (%): 1 -> 69 | Improved |
| add monotonic mixing network (QMIX vs VDN) | VDN · SMAC 27m_vs_30m | test win rate (%): 0 -> 49 | Improved |
| use QMIX instead of VDN | VDN · SMAC bane_vs_bane | test win rate (%): 94 -> 85 | Hurt |
| use QMIX instead of VDN | VDN · SMAC 3s_vs_5z | test win rate (%): 91 -> 87 | Hurt |
| add central state to VDN as a state-dependent bias (VDN-S) (ablation) | VDN · SMAC 3s5z, 2c_vs_64zg, MMM2 | test win rate | Improved |
| make mixing weights depend on state (QMIX vs QMIX-NS) (ablation) | QMIX-NS · SMAC 3s5z, 2c_vs_64zg, MMM2 | test win rate | Improved |
| add nonlinear final layer to mixing network (QMIX vs QMIX-Lin) (ablation) | QMIX-Lin · SMAC 3s5z, 2c_vs_64zg, MMM2 | test win rate | Mixed |
| use off-policy value-based factorisation instead of on-policy COMA | COMA · SMAC 14 scenarios | median test win rate | Improved |
| use QMIX instead of QTRAN | QTRAN · SMAC 3s5z and 2s3z | test win rate | Improved |
Closest papers by meaning
- QMIX: Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning (ICML 2018 · similarity 0.98)
- Weighted QMIX: Expanding Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning (NeurIPS 2020 · similarity 0.94)
- QTRAN: Learning to Factorize with Transformation for Cooperative Multi-Agent Reinforcement Learning (ICML 2019 · similarity 0.91)
- ConcaveQ: Non-monotonic Value Function Factorization via Concave Representations in Deep Multi-Agent Reinforcement Learning (AAAI 2024 · similarity 0.91)
- Towards Understanding Cooperative Multi-Agent Q-Learning with Value Factorization (NeurIPS 2021 · similarity 0.90)
- AgentMixer: Multi-Agent Correlated Policy Factorization (AAAI 2025 · similarity 0.88)
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.
