FAQ about the Flatland Repository
This section provides you with information about the most common questions around the Flatland repository. If your question is still not answered either reach out to the contacts listed on the repository directly or open an issue by following these guidlines.
How can I get started with Flatland?
Install Flatland by running pip install -U flatland-rl or directly from source by cloning the flatland repository and running python setup.py --install in the repository directory.
These Tutorials help you get a basic understanding of the flatland environment.
How do I train agents on Flatland?
Once you have installed Flatland, head over to the baselines repository to see how you can train your own reinforcement learning agent on Flatland.
Check out this tutorial to get a sense of how it works.
What is a observation builder and which should I use?
Observation builders give you the possibility to generate custom observations for your controller (reinfocement learning agent, optimization algorithm,...). The observation builder has access to all environment data and can perform any operations on them as long as they are not changed. This tutorial will give you a sense on how to use them.