This repository is part of a Semester project with SYCAMORE (Systems Control and Multiagent Optimization Research) Laboratory at EPFL. It contains implementations of log-linear learning and its modifications as part of validation of alogrithm proposed in [1].
The project was tested on Windows 11 and is written in Python 3.12. Other requirements include:
- numpy
- matplotlib
- pandas
- scipy
- networkx
- SciencePlots
Recommended:
- pytest (for running tests)
- Clone the repository:
git clone https://github.com/yourusername/potentialgames.git cd potentialgames - Install dependencies:
pip install -r [requirements.txt]
Scripts and modules for running experiments and plotting results are provided in the scripts/ and potentialgames/ directories. See the code and docstrings for details.
[1] Maddux, A., Ouhamma, R. and Kamgarpour, M., (2024). Finite time convergence to ε-efficient Nash Equilibirum in potential games. arXiv preprint arXiv:2405.15497.