Integrated Planning and Learning

Studying topics separately followed by learning about them together has been a recurring theme in this book. This chapter continues recombining different approaches. Specifically, the chapter combines the Model-based approaches and Model-free approaches to make the algorithms more powerful and sample efficient. This approach leverages the best of both of them and is the main emphasis of this chapter. You will also study the Exploration-exploitation dilemma in more detail to go beyond just blindly following e-greedy policies. You will look at simpler setups to gain a stronger understanding of the exploration-exploitation dilemma. This will be followed by a deep dive into a “guided forward looking tree search approach,” called Monte Carlo tree search (MCTS).