Credit Assignment with Meta-Policy Gradient for Multi-Agent Reinforcement Learning
This addresses a critical bottleneck in centralized training with decentralized execution for multi-agent systems, offering incremental improvements in performance.
The paper tackles the reward decomposition problem in multi-agent reinforcement learning by proposing a meta-learning-based mixing network with meta policy gradient (MNMPG) framework, which outperforms state-of-the-art algorithms on 4 out of 5 super hard StarCraft II scenarios.
Reward decomposition is a critical problem in centralized training with decentralized execution~(CTDE) paradigm for multi-agent reinforcement learning. To take full advantage of global information, which exploits the states from all agents and the related environment for decomposing Q values into individual credits, we propose a general meta-learning-based Mixing Network with Meta Policy Gradient~(MNMPG) framework to distill the global hierarchy for delicate reward decomposition. The excitation signal for learning global hierarchy is deduced from the episode reward difference between before and after "exercise updates" through the utility network. Our method is generally applicable to the CTDE method using a monotonic mixing network. Experiments on the StarCraft II micromanagement benchmark demonstrate that our method just with a simple utility network is able to outperform the current state-of-the-art MARL algorithms on 4 of 5 super hard scenarios. Better performance can be further achieved when combined with a role-based utility network.