Reinforcement Learning: Principles and Python Implementation

肖智清China Machine Press · ChinaNov 2025

4.5

Word count: 26K words
Chapters: 16

Summary

This book combines theory and practice. On the premise of ensuring theoretical completeness, it provides accompanying code for each chapter (especially deep reinforcement learning), and all accompanying code implementations are based on TensorFlow 2. In terms of theory, this book covers almost all mainstream classical reinforcement learning algorithms (such as eligibility traces) and deep reinforcement learning algorithms (such as deep deterministic policy gradient), which previous tutorials (whether Chinese or English, online or in electronic or print books) had not achieved.

Editor's review

The positioning of this book is clear: to put the theoretical framework of reinforcement learning and runnable code in the same book. It covers classical reinforcement learning algorithms and deep reinforcement learning algorithms, and provides TensorFlow 2-based implementations for the deep reinforcement learning part, which is especially important for self-learners—many textbooks stop after explaining the formulas, and readers still do not know how to put them into practice. While ensuring theoretical completeness, this book emphasizes code practice and is suitable as a bridge from "understanding the concepts" to "hands-on implementation." When reading, it is recommended not to skip the formula derivations, nor to only read code without writing code; it is best to run the examples while reading, then try changing parameters and observing the results. For learners who need to systematically sort out the algorithm landscape and also want engineering footholds, this is a highly practical reference book.

Who it's for

A good fit for

Readers with some machine learning background who want to systematically learn reinforcement learning
Upper-year students who need to implement algorithms hands-on with TensorFlow 2
Engineers who want to complete their map of classical and deep reinforcement learning algorithms
People doing course projects or introductory research who need both theory and code references

Not a good fit for

Complete beginners with no programming or machine learning background
Readers who only want to understand reinforcement learning concepts and do not plan to write code
Readers who prefer pure mathematical theory derivation and do not need engineering implementation

Reading tips

First read the early chapters to establish basic concepts such as Markov decision processes, then move on to the algorithm chapters.
Run the code in each chapter by hand at least once, change key parameters, and record how the results change.
Read classical algorithms alongside their corresponding deep algorithms, and organize your own notes on the algorithm landscape.

About the publisher

China Machine Press China · Founded in 1950

In addition to science and technology and professional publishing, the books in our catalog from China Machine Press also reflect a sustained focus on psychology, computing, management, finance and economics, and social sciences. In psychology, the catalog includes works by C. G. Jung and introductions to analytical psychology, as well as applied topics such as stress and health, and addiction and healing. In computing, it features Structure and Interpretation of Computer Programs (JavaScript Edition). Management and finance titles include Administrative Behavior (4th Edition), The Price of Inequality (Collector's Edition), and This Time Is Different: Eight Centuries of Financial Folly (Commemorative Edition). In the philosophy and history of science, there is The Worldview: The Third Edition. Relevant authors include C. G. Jung, William McGuire, Sonu Shamdasani, and Gabor Maté. Public information is limited; the above summarizes the publisher's profile based solely on the books in our catalog.

版权信息
前言
第1章初识强化学习
1.1.强化学习及其关键元素
1.2.强化学习的应用
1.3.智能体/环境接口
1.4.强化学习的分类
1.5.如何学习强化学习
1.6.案例：基于Gym库的智能体/环境交互
1.7.本章小结
第3章有模型数值迭代
3.1.有模型策略迭代
3.2.有模型价值迭代
3.3.动态规划
3.4.案例：冰面滑行
3.5.本章小结

FAQ

What background is needed to read this book?

Basic knowledge of linear algebra and probability theory is needed, as well as experience with Python programming; a preliminary understanding of machine learning, especially neural networks, will make reading smoother. The theoretical part of the book is fairly complete, but readers with no programming background at all will find it difficult.

What framework is used to implement the code in the book?

According to the available information, the accompanying code for the deep reinforcement learning part of the book is implemented based on TensorFlow 2. Readers are advised to become familiar with the basics of TensorFlow 2 in advance and to run the code while reading for better results.

What algorithms does this book cover?

The available information shows that this book covers almost all mainstream classical reinforcement learning algorithms (such as eligibility traces) and deep reinforcement learning algorithms (such as deep deterministic policy gradient), and provides accompanying code for each chapter, especially the deep reinforcement learning part, giving equal weight to theory and implementation.

Comments

Find this book

This site is an index: you start “Search all sources” in your browser, and the entries it lists all point to third-party sites; please judge their availability yourself.

About the author

肖智清 Author

Xiao Zhiqing's publicly available information is limited. At present, only one work by this author is known, namely Reinforcement Learning: Principles and Python Implementation, commonly classified under programming and design. Judging from the title and category, the writing centers on reinforcement learning, a branch of machine learning, combining explanations of theoretical principles with code implementation in Python, which places the book in the field of practice-oriented technical writing. Such subject matter generally requires the author to be familiar with both the mathematics and logic behind algorithms and the translation of abstract concepts into runnable code, and thus expects a certain level of hands-on ability from readers. Although no personal details such as biography, education, or professional background are recorded in the database and cannot be introduced further, the book's choice of topic suggests that the author focuses on turning cutting-edge methods into practical learning materials, with a style that gives equal weight to principles and implementation and uses code to aid understanding, making it suitable for readers who prefer to learn by doing.