Reinforcement Learning, second edition: An Introduction Adaptive Computation and Machine Learning series 2nd Edition

One of the most active study topics in artificial intelligence is reinforcement learning, which is a computational method to learning in which an agent strives to maximize the overall amount of reward it receives while interacting with a complicated, unpredictable environment. Richard Sutton and Andrew Barto’s Reinforcement Learning provides a clear and basic introduction of the field’s core theories and techniques. This second version has been extensively enlarged and updated, with new topics included and existing ones updated.
This second edition, like the first, concentrates on key online learning algorithms, with extra mathematical information separated out in colored boxes. Part I includes as much reinforcement learning as feasible without going beyond the tabular case with exact solutions. There are several algorithms.