Reinforcement Learning for Sequential Decision and Optimal Control

Reinforcement learning (RL) is a theory that enables an agent to develop the self-evolution ability through continuing environment interactions. It has been successful in various fields, such as chess games, computer games, and robotic control, and is considered a promising and powerful tool to create general artificial intelligence in the future. This book aims to help researchers and practitioners take a comprehensive view of RL and understand the in-depth connection between RL and optimal control. It includes systematic and thorough explanations of theoretical basics, as well as methodical guidance of practical algorithm implementations. The book covers classic theories and recent achievements, and is carefully and logically organized.

Format: Hardback
Length: 462 pages
Publication date: 01 December 2022
Publisher: Springer Verlag, Singapore

Reinforcement learning (RL) is a theory that enables an agent to develop the self-evolution ability through continuing environment interactions. It has been successful in various fields, such as chess games, computer games, and robotic control, and is considered a promising and powerful tool to create general artificial intelligence in the future. The book aims to help researchers and practitioners understand the inherent mechanism of reinforcement learning, the internal connection between RL and optimal control, the evolution of RL in the past few decades, the practical and effective RL algorithms for real-world scenarios, the key challenges that RL faces today, and the current trend of RL research.

The book begins by introducing the concept of reinforcement learning and its applications in different fields. It then discusses the underlying mechanism of reinforcement learning, including the definition of reinforcement, the reinforcement signal, the reward function, and the policy iteration process. The book also explores the internal connection between reinforcement learning and optimal control, including the use of dynamic programming, stochastic optimization, and reinforcement learning algorithms.

The book then discusses the evolution of reinforcement learning in the past few decades, including the development of deep learning, transfer learning, and multi-agent reinforcement learning. It also highlights the milestones in the development of reinforcement learning, such as the introduction of the Monte Carlo method, the development of Q-learning algorithms, and the introduction of the actor-critic architecture.

The book also discusses the practical and effective RL algorithms for real-world scenarios, including the use of deep reinforcement learning, policy search algorithms, and reinforcement learning with human supervision. It also addresses the key challenges that RL faces today, such as the exploration of complex environments, the stability of learning algorithms, and the generalization of learning to new tasks.

The book also discusses the current trend of RL research, including the use of RL in autonomous driving, healthcare, finance, and other fields. It also highlights the open research questions in RL, such as the development of more efficient RL algorithms, the exploration of new reinforcement learning paradigms, and the integration of RL with other AI technologies.

In conclusion, the book provides a comprehensive and up-to-date introduction to reinforcement learning and its applications in different fields. It is a valuable resource for researchers, practitioners, and students interested in learning more about this interdisciplinary field of trial-and-error learning and optimal control.

Weight: 994g
Dimension: 240 x 168 (mm)
ISBN-13: 9789811977831
Edition number: 1st ed. 2023