Research map / Multi-agent systems and games / Cooperative multi-agent RL
Efficient Communication in Multi-Agent Reinforcement Learning via Variance Based Control
Sai Qian Zhang, Qi Zhang, Jieyu Lin
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 variance-controlled, confidence-triggered communication instead of SchedNet scheduling | VBC on SMAC · MMM | communication overhead: 5.25% vs 50% (VBC+VDN); 5.36% vs 50% (VBC+QMIX) | Improved |
| Use variance-controlled, confidence-triggered communication instead of SchedNet scheduling | VBC on SMAC · 2s3z | communication overhead: 4.33% vs 60% (VBC+VDN); 4.68% vs 60% (VBC+QMIX) | Improved |
| Use variance-controlled, confidence-triggered communication instead of SchedNet scheduling | VBC on SMAC · 3s5z | communication overhead: 27.70% vs 62.5% (VBC+VDN); 28.13% vs 62.5% (VBC+QMIX) | Improved |
| Use variance-controlled, confidence-triggered communication instead of SchedNet scheduling | VBC on SMAC · 3s_vs_4z | communication overhead: 5.07% vs 33.3% (VBC+VDN); 5.19% vs 33.3% (VBC+QMIX) | Improved |
| Use variance-controlled, confidence-triggered communication instead of SchedNet scheduling | VBC on SMAC · 6h_vs_8z | communication overhead: 35.93% vs 66.7% (VBC+VDN); 36.16% vs 66.7% (VBC+QMIX) | Improved |
| Use variance-controlled, confidence-triggered communication instead of SchedNet scheduling | VBC on SMAC · 6z_vs_24zerg | communication overhead: 12.13% vs 50% (VBC+VDN); 13.35% vs 50% (VBC+QMIX) | Improved |
| Add variance penalty and execution-time variance filtering to full communication (ablation) | VBC+VDN / FC · SMAC battle tasks | winning rate | Mixed |
| Use VBC rather than fixed-number Top(k) scheduling | VBC vs SchedNet · SMAC battle tasks | winning rate | Improved |
| Enable communication instead of using communication-free value decomposition | VBC, SchedNet, and FC vs VDN/QMIX · SMAC battle tasks | winning rate | Improved |
| Use VBC communication | VBC · SMAC | winning rate: 20% higher | Improved |
Closest papers by meaning
- Adaptive Communication Range for Scalable Cooperative Multi-Agent Reinforcement Learning (NeurIPS 2026 · similarity 0.90)
- Multi-Agent Incentive Communication via Decentralized Teammate Modeling (AAAI 2022 · similarity 0.89)
- Efficient Communication in Multi-Agent Reinforcement Learning with Implicit Consensus Generation (AAAI 2025 · similarity 0.89)
- Multi-Agent Reinforcement Learning with Communication-Constrained Priors (NeurIPS 2025 · similarity 0.89)
- Exponential Topology-enabled Scalable Communication in Multi-agent Reinforcement Learning (ICLR 2025 · similarity 0.88)
- Robust Communicative Multi-Agent Reinforcement Learning with Active Defense (AAAI 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.
