MaMuJoCo

The 'MaMuJoCo' dataset/benchmark is used to evaluate offline multi-agent reinforcement learning (MARL) algorithms by providing a set of environments that simulate multi-agent interactions and their corresponding joint action spaces.

Papers using MaMuJoCo (11)

ACPO: Agent-Chained Policy Optimization for Multi-Agent Reinforcement Learning2026CoFlow: Coordinated Few-Step Flow for Offline Multi-Agent Decision Making2026CODA: Coordination via On-Policy Diffusion for Multi-Agent Offline Reinforcement Learning2026STAIRS-Former: Spatio-Temporal Attention with Interleaved Recursive Structure Transformer for Offline Multi-task Multi-agent Reinforcement Learning2026Learning Partial Action Replacement in Offline MARL2026Prism: Spectral Parameter Sharing for Multi-Agent Reinforcement Learning2026Diffusing to Coordinate: Efficient Online Multi-Agent Diffusion Policies2026Preference-Guided Learning for Sparse-Reward Multi-Agent Reinforcement Learning2025Revisiting Multi-Agent World Modeling from a Diffusion-Inspired Perspective2025O-MAPL: Offline Multi-agent Preference Learning2025RMIO: A Model-Based MARL Framework for Scenarios with Observation Loss in Some Agents2024