Timezone: »

MASER: Multi-Agent Reinforcement Learning with Subgoals Generated from Experience Replay Buffer
JEON JEEWON · WOOJUN KIM · Whiyoung Jung · Youngchul Sung

Wed Jul 20 03:30 PM -- 05:30 PM (PDT) @ Hall E #825

In this paper, we consider cooperative multi-agent reinforcement learning (MARL) with sparse reward. To tackle this problem, we propose a novel method named MASER: MARL with subgoals generated from experience replay buffer. Under the widely-used assumption of centralized training with decentralized execution and consistent Q-value decomposition for MARL, MASER automatically generates proper subgoals for multiple agents from the experience replay buffer by considering both individual Q-value and total Q-value. Then, MASER designs individual intrinsic reward for each agent based on actionable representation relevant to Q-learning so that the agents reach their sub-goals while maximizing the joint action value. Numerical results show that MASER significantly outperforms StarCraft II micromanagement benchmark compared to other state-of-the-art MARL algorithms.

Author Information

JEON JEEWON (Korea Advanced Institute of Science and Technology)
Whiyoung Jung (KAIST)
Youngchul Sung (KAIST)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors