https://github.com/aleksandarhaber/greedy-in-the-limit-with-infinite-exploration-glie-monte-carlo-reinforcement-learning-in-python