Thank you very much indeed for this tutorial. May I ask, is it possible to use Q-learning (Part 0) to do the same thing as the policy agents here? Did you try to make a comparison? Thanks again.