Deep Reinforcement Learning Hands-On
上QQ阅读APP看书,第一时间看更新

Chapter 5. Tabular Learning and the Bellman Equation

In the previous chapter, we got acquainted with our first Reinforcement Learning (RL) method, cross-entropy, and saw its strengths and weaknesses. In this new part of the book, we'll look at another group of methods, called Q-learning, which have much more flexibility and power.

This chapter will establish the required background shared by those methods. We'll also revisit the FrozenLake environment and show how new concepts will fit with this environment and help us to address the issues of the environment's uncertainty.