The author proposes the Imagine, Initialize, and Explore (IIE) method to enhance multi-agent exploration by leveraging a transformer model for trajectory generation. This method outperforms existing approaches on various benchmarks.
IIE proposes a novel method for efficient multi-agent exploration in complex scenarios using a transformer model to imagine critical states and trajectories before exploration.
MAexp is a generic high-efficiency platform that integrates a broad range of state-of-the-art multi-agent reinforcement learning (MARL) algorithms and representative exploration scenarios to enable rigorous evaluation and comparison of multi-agent exploration techniques.