simple Q-learning algorithm to explore a maze overview Let's create very simple Q-learning agent that explore an arbitrary-sized maze.