Hydrogym Reference Documentation

Hydrogym is an open-source library of challenge problems in data-driven modeling and control of fluid dynamics for state-of-the-art reinforcement learning algorithms. For that purposes it exposes a number of APIs at multiple levels to allow every user to interact with the environment in the way most befitting of their algorithmic approach. While the current

Core Features

Hierarchical Design

Hierarchically designed for clean, industry-standard APIs with the expanded hydrogym.env.FlowEnv class extending the gymnasium.Env interface.

Scalability

Readily scalable experimentation infrastructure with Ray to support the rapid experimentation at cluster-scale.

4 Environments

Provision of four environments for flow control with reinforcement learning with varying computational requirements.

RL Interfaces

Reinforcement learning interfaces to a broad set of the most popular RL frameworks used today.

Note

While there exists a pip-install, this project is still under heavy active development with breaking changes happening with little to no lead time.