Algorithmic Foundations of Reinforcement Learning

A comprehensive algorithmic introduction to reinforcement learning is given, laying the foundational concepts and methodologies. Fundamentals of Markov Decision Processes (MDPs) and dynamic programming are covered, describing the principles and techniques for addressing model-based problems within MDP frameworks. The most significant model-free reinforcement learning algorithms, including Q-learning and actor-critic methods are explained in detail. A comprehensive overview of each algorithm’s mechanisms is provided, forming a robust algorithmic and mathematical understanding of current practices in reinforcement learning.