Reinforcement Learning An introduction

by Richard Sutton and Andrew Barto

http://incompleteideas.net/book/the-book-2nd.html