Discretizing Continuous Action Space for On-Policy Optimization

Abstract

In this work, we show that discretizing action space for continuous control is a simple yet powerful technique for on-policy optimization. The explosion in the number of discrete actions can be efficiently addressed by a policy with factorized distribution across action dimensions. We show that the discrete policy achieves significant performance gains with state-of-the-art on-policy optimization algorithms (PPO, TRPO, ACKTR) especially on high-dimensional tasks with complex dynamics. Additionally, we show that an ordinal parameterization of the discrete distribution can introduce the inductive bias that encodes the natural ordering between discrete actions. This ordinal architecture further significantly improves the performance of PPO/TRPO.

Cite

Text

Tang and Agrawal. "Discretizing Continuous Action Space for On-Policy Optimization." AAAI Conference on Artificial Intelligence, 2020. doi:10.1609/AAAI.V34I04.6059

Markdown

[Tang and Agrawal. "Discretizing Continuous Action Space for On-Policy Optimization." AAAI Conference on Artificial Intelligence, 2020.](https://mlanthology.org/aaai/2020/tang2020aaai-discretizing/) doi:10.1609/AAAI.V34I04.6059

BibTeX

@inproceedings{tang2020aaai-discretizing,
  title     = {{Discretizing Continuous Action Space for On-Policy Optimization}},
  author    = {Tang, Yunhao and Agrawal, Shipra},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year      = {2020},
  pages     = {5981-5988},
  doi       = {10.1609/AAAI.V34I04.6059},
  url       = {https://mlanthology.org/aaai/2020/tang2020aaai-discretizing/}
}