The book that defined how a generation learned reinforcement learning — significantly expanded and updated in its second edition.

Sutton and Barto’s Reinforcement Learning: An Introduction gives a clear, unhurried account of how an agent learns to act by maximising reward while interacting with a complex, uncertain environment. Part I develops the tabular case, where exact solutions exist; Part II moves to function approximation and the methods that scale to real problems. The heavier mathematics sits in shaded boxes, so the main thread stays readable for anyone with a working grasp of probability.