mattmolinare/correlated-q

Correlated-Q Learning

Learn equilibrium strategies for the two-player game RoboCub with the following reinforcement learning algorithms:

• Q-learning
• Foe-Q learning
• Friend-Q learning
• Correlated-Q learning

Setting up the environment

To create the environment

conda env create -f environment.yml

To activate the environment

conda activate ceq

Running the code

To generate Q-values using default Correlated-Q learning parameters

python main.py configs/ceq_learner.py results