ABOUT

This is a reinforcement learning implementation to generate a schedule for
a sports season. It was originally implemented as an alternative to the
algorithm for powerplay (see
https://github.com/jak103/powerplay/tree/61d4d7ed41539dee7c261418b28624a9ef0d63d5/backend/internal/server/apis/schedule
for the golang implementation). The main goals of the reinforcement model are to
schedule games such that teams have a sufficient number of days between games,
and also to try to balance the time of day that teams play (i.e. try to prevent
a team from having all their games at 22:30).


STATUS

The reinforcement model is working, however the schedules it is generating do
not seem to be as good as the golang implementation, so development on this
project was halted in favor of that implementation linked above.


REQUIREMENTS

- python 3.3+


GETTING STARTED

python -m venv .env
source .env/bin/activate  # assuming using bash or zsh
pip install -r requirements.txt
python model.py


LICENSE

This is licensed under the permissive MIT license. See LICENSE for details.


CONTACT

To contact the author of this project please email me:
gl@marcusquincy.org