From General Relation Patterns to Task-Specific Decision-Making in Continual Multi-Agent Coordination
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866916832622936064 |
|---|---|
| author | Yao, Chang Lin, Youfang Song, Shoucheng Wu, Hao Ma, Yuqing Han, Shang Lv, Kai |
| author_facet | Yao, Chang Lin, Youfang Song, Shoucheng Wu, Hao Ma, Yuqing Han, Shang Lv, Kai |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_06004 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | 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, Shang Lv, Kai Multiagent Systems 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. |
| title | From General Relation Patterns to Task-Specific Decision-Making in Continual Multi-Agent Coordination |
| topic | Multiagent Systems |
| url | https://arxiv.org/abs/2507.06004 |