.png){kind=icon}
.png){kind=icon}
Return
ROCO: Role-oriented communication for efficient multi-agent reinforcement learning
.png){kind=icon}
PRE
AI
DOI：10.1016/j.eswa.2025.129421.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Dual-layer framework minimizes redundancy in multi-agent communication. • Adaptive roles assigned using mutual information for dynamic task adaptation. • Attention-based filtering enhances bandwidth efficiency and coordination. • Outperforms state-of-the-art in SMAC and GRF with faster convergence, higher win rate.
Keywords:
multi-agent communication
adaptive roles
mutual information
attention-based filtering
bandwidth efficiency
Journal
IF:
7.5
Papers:
2.9W
Citations:
10.2W