Book Image

Python Reinforcement Learning

By : Sudharsan Ravichandiran, Sean Saito, Rajalingappaa Shanmugamani, Yang Wenzhuo
Book Image

Python Reinforcement Learning

By: Sudharsan Ravichandiran, Sean Saito, Rajalingappaa Shanmugamani, Yang Wenzhuo

Overview of this book

Reinforcement Learning (RL) is the trending and most promising branch of artificial intelligence. This Learning Path will help you master not only the basic reinforcement learning algorithms but also the advanced deep reinforcement learning algorithms. The Learning Path starts with an introduction to RL followed by OpenAI Gym, and TensorFlow. You will then explore various RL algorithms, such as Markov Decision Process, Monte Carlo methods, and dynamic programming, including value and policy iteration. You'll also work on various datasets including image, text, and video. This example-rich guide will introduce you to deep RL algorithms, such as Dueling DQN, DRQN, A3C, PPO, and TRPO. You will gain experience in several domains, including gaming, image processing, and physical simulations. You'll explore TensorFlow and OpenAI Gym to implement algorithms that also predict stock prices, generate natural language, and even build other neural networks. You will also learn about imagination-augmented agents, learning from human preference, DQfD, HER, and many of the recent advancements in RL. By the end of the Learning Path, you will have all the knowledge and experience needed to implement RL and deep RL in your projects, and you enter the world of artificial intelligence to solve various real-life problems. This Learning Path includes content from the following Packt products: • Hands-On Reinforcement Learning with Python by Sudharsan Ravichandiran • Python Reinforcement Learning Projects by Sean Saito, Yang Wenzhuo, and Rajalingappaa Shanmugamani
Table of Contents (27 chapters)
Title Page
About Packt
Contributors
Preface
Index

The Bellman equation and optimality


The Bellman equation, named after Richard Bellman, American mathematician, helps us to solve MDP. It is omnipresent in RL. When we say solve the MDP, it actually means finding the optimal policies and value functions. There can be many different value functions according to different policies. The optimal value function 

 is the one which yields maximum value compared to all the other value functions:

 

Similarly, the optimal policy is the one which results in an optimal value function.

Since the optimal value function 

is the one that has a higher value compared to all other value functions (that is, maximum return), it will be the maximum of the Q function. So, the optimal value function can easily be computed by taking the maximum of the Q function as follows:

  -- (3)

The Bellman equation for the value function can be represented as, (we will see how we derived this equation in the next topic):

It indicates the recursive relation between a value of a state...