A short review and comparison of Q-Learning, Function Approximation by gradient descent and Monte Carlo Tree Search algorithms, implemented to run on an environment based on an emulation of the Nintendo Entertainment System ...