The Art of Reinforcement Learning: Fundamentals, Mathematics, and Implementations with Python

© The Author(s), under exclusive license to APress Media, LLC, part of Springer Nature 2023

M. HuThe Art of Reinforcement Learninghttps://doi.org/10.1007/978-1-4842-9606-6_14

14. Planning with a Model: AlphaZero

Michael Hu¹  

(1)

Shanghai, Shanghai, China

 

Throughout this book, we have discussed various model-free reinforcement learning algorithms that have proven effective in solving simple reinforcement learning problems such as classic control tasks and Atari video games. However, for more complex problems, these algorithms may not perform well, even with the application of advanced techniques like distributed training and curiosity-driven exploration. Therefore, in this final chapter, we will explore how to use model-based reinforcement learning ...