Multi armed bandit
An implementation of a range of epsilon-greedy algorithms to solve the multi-armed bandits problem.
Install
- Clone the repo locally:
git clone git@github.com:SebastianoF/multi-armed-bandits-testbed.git
cd multi-armed-bandits-testbed
- Create and activate a virtual environment:
virtualenv -p python3.9 venv
source venv/bin/activate
The library is tested with python 3.9, 3.10, 3.11. Could work with previous version downgrading manually numpy to a compatible version.
- Install in development mode:
pip install -e .
- To install the development third party requirement packages along with the basic ones:
pip install -e .["dev"]
Where to start
Check out the examples in the folder /examples/start_here.py.
Development
- Dependent libraries are managed with pip-compile-multi
- Continuous integration is integrated with CircleCI
- Code formatting happens via makefile command
make reformat
, after installing the dev packages.
Resources
-
R. Sutton, A. Barto, "Reinforcement Learning, an introduction", Chapter 1.
-
Further introduction can be found in the folder
/docs/
Licence
Repository open-sourced under MIT licence.