Deep Multi-Agent Reinforcement Learning with Discrete-Continuous Hybrid Action Spaces
This solves a practical problem for multi-agent reinforcement learning researchers and practitioners by enabling effective handling of hybrid action spaces, though it is incremental in extending existing paradigms to this specific case.
The paper tackles the problem of applying deep reinforcement learning to multi-agent systems with discrete-continuous hybrid action spaces, which had not been addressed before, and shows that their proposed algorithms (Deep MAPQN and Deep MAHHQN) significantly outperform existing methods on tasks like RoboCup Soccer and Ghost Story.
Deep Reinforcement Learning (DRL) has been applied to address a variety of cooperative multi-agent problems with either discrete action spaces or continuous action spaces. However, to the best of our knowledge, no previous work has ever succeeded in applying DRL to multi-agent problems with discrete-continuous hybrid (or parameterized) action spaces which is very common in practice. Our work fills this gap by proposing two novel algorithms: Deep Multi-Agent Parameterized Q-Networks (Deep MAPQN) and Deep Multi-Agent Hierarchical Hybrid Q-Networks (Deep MAHHQN). We follow the centralized training but decentralized execution paradigm: different levels of communication between different agents are used to facilitate the training process, while each agent executes its policy independently based on local observations during execution. Our empirical results on several challenging tasks (simulated RoboCup Soccer and game Ghost Story) show that both Deep MAPQN and Deep MAHHQN are effective and significantly outperform existing independent deep parameterized Q-learning method.