Welcome to Flatland

Flatland is simple and efficient grid world for multi-agent reinfrocement learning tasks. It allows for a diverse set of problem classes to be implemented from free moving agents to restricted movement on transportation networks.

How to get started

To get started with the Flatland environment follow the steps in installation and look at the examples.

Specifications

• Core
• Core
• Rendering
• Environment
• Railway

Baselines

• Single Agent Navigation
• Multi Agent Navigation

Integrations Tests

• Tests

Links and Files

• Links
• Files