RetaiL is a library aimed at improving experimentation in a grocery store context, to reduce generated waste. We think it is a great addition for both reinforcement learning and supply chain researchers, but also to data scientists working in retail environments.
We design RetaiL around two main components:
- Item Generation contains the item Generation, scripted in R.
- Store environment
We additionally provide three usage examples of RetaiL:
- Transporation cost impact on marginal ordering cost
- Daily customer distribution impact on ordering policy performance
- Risk assesment of an ordering policy
RetaiL requires R, with the copula package to generate items. Moreover, RetaiL requires rlpyt in its version 0.1.1.dev0. RetaiL requires Python 3.6 or 3.7.
# Checkout this git repository
# Create a virtual environment
python3 -m venv env
# Activate the environment
source env/bin/activate
# Install the dependencies
python3 setup.py install
Run server:
python3 app.py
View the grocery store simulation in your web browser at http://localhost:8050/.
A development environment can be set up regardless of the developer's operating system by using Docker.
docker build . -t retail
docker run -ti --rm -p 8050:8050 retail