RLlib: Abstractions for Distributed Reinforcement Learning

Eric Liang

Reinforcement learning (RL) algorithms involve the deep nesting of highly irregular computation patterns, each of which typically exhibits opportunities for distributed computation. We argue for distributing RL components in a composable way by adapting algorithms for top-down hierarchical control, thereby encapsulating parallelism and resource requirements within short-running compute tasks. We demonstrate the benefits of this principle through RLlib: a library that provides scalable software primitives for RL. These primitives enable a broad range of algorithms to be implemented with high performance, scalability, and substantial code reuse. RLlib is available at https://github.com/ray-project/ray.

Published On: December 7, 2017

Presented At/In: International Conference on Machine Learning (ICML 2018)

Link: https://arxiv.org/abs/1712.09381

Authors: Eric Liang, Richard Liaw, Robert Nishihara, Philipp Moritz, Roy Fox, Kenneth Goldberg, Joseph Gonzalez, Michael Jordan, Ion Stoica