This is an educational project to see the inner workings of the Nash-Q Learning algorithm. The Nash-Q Learning algorithm is a multi-agent reinforcement learning algorithm that is designed to learn Nash equilibria in general-sum stochastic games. This project is designed to be educational and is not intended to be used in production environments.
The main entry point for the project is the Application.ipynb notebook. There you can specify a Stochastic Game and run the Nash-Q Learning algorithm on it. To use the notebook you can just download the file and run it.
We advice you to use the VSCode IDE with the Jupyter extension to run the notebook optimally.
❗**Note:**❗: A Colab version has been developed and can be found at the following link. Unfortunately, due to widget compatibility issues, the notebook may work unexpectedly on Google Colab, we strongly suggest downloading the local version. https://colab.research.google.com/drive/1-EwH_ONeIwtgkNf7MFr0kPj8RjwuUHg4?usp=sharing
If you are curious about the algorithm implementation you can look at ./LearningNashQLearning/LearningNashQLearning/Model/NashQLearning.py.