We present a framework combining hierarchical and multi-agent deep reinforcement learning approaches to solve coordination problems among a multitude of agents using a semi-decentralized model. The framework extends the multi-agent learning setup by introducing a meta-controller that guides the communication between agent pairs, enabling agents to focus on communicating with only one other agent at any step. This hierarchical decomposition of the task allows for efficient exploration to learn policies that identify globally optimal solutions even as the number of collaborating agents increases. We show promising initial experimental results on a simulated distributed scheduling problem.
@article{arxiv.1712.08266,
title = {Federated Control with Hierarchical Multi-Agent Deep Reinforcement Learning},
author = {Saurabh Kumar and Pararth Shah and Dilek Hakkani-Tur and Larry Heck},
journal= {arXiv preprint arXiv:1712.08266},
year = {2017}
}
Comments
Hierarchical Reinforcement Learning Workshop at the 31st Conference on Neural Information Processing Systems