Welcome to gym-DSSAT documentation!#
gym-DSSAT is a modification of the Decision Support System for Agrotechnology Transfer (DSSAT) Fortran software into an easy to manipulate Python Open AI gym environment for Reinforcement Learning (RL) researchers. gym-DSSAT allows daily based interactions during the growing season between an RL agent and the crop model with usual gym conventions. gym-DSSAT is proudly powered by the PDI Data Interface (PDI) !
Attention
At the moment gym-DSSAT only supports most common Linux distributions! If your distribution is not supported, you can still check our Docker images, see installation Using Docker.
Companion article#
Please have a look to the companion paper which provides complementary information about gym-DSSAT, including use cases.
Citing gym-DSSAT#
@phdthesis{gautron2022gym,
title={gym-DSSAT: a crop model turned into a Reinforcement Learning environment},
author={Gautron, Romain and Gonzalez, Emilio Jos{\'e} Padron and Preux, Philippe and Bigot, Julien and Maillard, Odalric-Ambrym and Emukpere, David},
year={2022},
school={Inria Lille}
}
Help#
Any question/remark/request? The main support channel for gym-DSSAT is our mailing list.
Before asking a question
Please check if you can find an answer in the archives of our mail-list.
Contributing#
Any new issue is welcomed! For the most motivated, we have a todo list!
Acknowledgements#
We acknowledge the DSSAT team, especially Gerrit Hoogenboom and Cheryl Porter. Thanks to the PDI team, especially to Julien Bigot. We acknowledge Bruno Raffin, leader of Inria DataMove team, for his support. We acknowledge the Consultative Group for International Agricultural Research (CGIAR) Platform for Big Data in Agriculture, special thanks to Brian King. Thanks to the French Agricultural Research Centre for International Development (CIRAD) and the French Institute for Research in Computer Science and Automation (Inria), in particular the SCOOL team, for their support.