Skip to content

Latest commit

 

History

History
20 lines (12 loc) · 1 KB

README.md

File metadata and controls

20 lines (12 loc) · 1 KB
Raw

code style

This project is a work in progress, we want to learn the PDDL domain of Starcraft. If you work on a numeric multi-agent solver for PDDL or want to contribute to this project, please feel free to connect us.

Getting Started

First, you will need to install StarCraft II. On windows or mac, follow the instructions on the StarCraft website. For linux, you can use the bash script here.

Then simply install SMAC as a package:

pip install git+https://github.com/oxwhirl/smacv2.git

For more additional information please see the original oxwhirl/smacv2 reposetory.

Usage

We use the gym_pddl.py to translate the gym actions to a PDDL game trace (see solutions/pfile0.trajectory). You can run main.py to see an example of play and a new trace generated.