Chapter 9. Policy Gradients – An Alternative

In this first chapter of part three of the book, we’ll consider an alternative way to handle Markov Decision Process (MDP) problems, which forms a full family of methods called Policy Gradients (PG). The chapter will present an overview of the methods, their motivation, and their strengths and weaknesses in comparison to the already familiar Q-learning. We will start with a simple PG method called REINFORCE and will try to apply it to our CartPole environment, comparing this with the Deep Q-Networks (DQN) approach.

Values and policy

Before we start talking about (PG), let’s refresh our minds with the common characteristics of the methods covered in part two of this book. The central topic in Q-learning ...

Get Deep Reinforcement Learning Hands-On now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.