From General Relation Patterns to Task-Specific Decision-Making in Continual Multi-Agent Coordination

Yao, Chang; Lin, Youfang; Song, Shoucheng; Wu, Hao; Ma, Yuqing; Han, Sheng; Lv, Kai

doi:10.24963/ijcai.2025/759

From General Relation Patterns to Task-Specific Decision-Making in Continual Multi-Agent Coordination

Chang Yao, Youfang Lin, Shoucheng Song, Hao Wu, Yuqing Ma, Sheng Han, Kai Lv

Download BibTeX Cite

Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence. Main Track. Pages 6821-6829.

https://doi.org/10.24963/ijcai.2025/759

Keywords

Agent-based and Multi-agent Systems: MAS: Coordination and cooperation
Machine Learning: ML: Multiagent Reinforcement Learning

Abstract

Continual Multi-Agent Reinforcement Learning (Co-MARL) requires agents to address catastrophic forgetting issues while learning new coordination policies with the dynamics team. In this paper, we delve into the core of Co-MARL, namely Relation Patterns, which refer to agents’ general understanding of interactions. In addition to generality, relation patterns exhibit task-specificity when mapped to different action spaces. To this end, we propose a novel method called General Relation Patterns-Guided Task-specific Decision-Maker (RPG). In RPG, agents extract relation patterns from dynamic observation spaces using a relation capturer. These task-agnostic relation patterns are then mapped to different action spaces via a task-specific decision-maker generated by a conditional hypernetwork. To combat forgetting, we further introduce regularization items on both the relation capturer and the conditional hypernetwork. Results on SMAC and LBF demonstrate that RPG effectively prevents catastrophic forgetting when learning new tasks and achieves zero-shot generalization to unseen tasks.

Chang Yao, Youfang Lin, Shoucheng Song, Hao Wu, Yuqing Ma, Sheng Han, and Kai Lv. 2025. From General Relation Patterns to Task-Specific Decision-Making in Continual Multi-Agent Coordination. In Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence, IJCAI-25. IJCAI, 6821–6829. https://doi.org/10.24963/ijcai.2025/759

@inproceedings{ijcai2025p759,
  title     = {From General Relation Patterns to Task-Specific Decision-Making in Continual Multi-Agent Coordination},
  author    = {Yao, Chang and Lin, Youfang and Song, Shoucheng and Wu, Hao and Ma, Yuqing and Han, Sheng and Lv, Kai},
  booktitle = {Proceedings of the Thirty-Fourth International Joint Conference on
               Artificial Intelligence, {IJCAI-25}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {James Kwok},
  pages     = {6821--6829},
  year      = {2025},
  month     = {8},
  note      = {Main Track},
  doi       = {10.24963/ijcai.2025/759},
  url       = {https://doi.org/10.24963/ijcai.2025/759},
}

TY  - CPAPER
TI  - From General Relation Patterns to Task-Specific Decision-Making in Continual Multi-Agent Coordination
AU  - Yao, Chang
AU  - Lin, Youfang
AU  - Song, Shoucheng
AU  - Wu, Hao
AU  - Ma, Yuqing
AU  - Han, Sheng
AU  - Lv, Kai
T2  - Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence, IJCAI-25
PB  - International Joint Conferences on Artificial Intelligence Organization
PY  - 2025
SP  - 6821
EP  - 6829
DO  - 10.24963/ijcai.2025/759
UR  - https://doi.org/10.24963/ijcai.2025/759
ER  -