This course is an introduction to models and algorithms for stochastic sequential decision-making under uncertainty. These problems arise in many important domains, ranging from robotic control, revenue management, to Internet advertising optimization, and personalized medicine and clinical trials. The goal of this course is to endow the student with a) a solid understanding of the foundation of Markov chains, Markov decision processes and reinforcement learning; and b) the ability to apply reinforcement learning algorithms to solve online decision-making problems in the context of diverse engineering and science problems.
Topics covered will include a review of probability (random variables, expectation, variance, etc.), reinforcement learning, Markov decision process, dynamic programming (e.g., policy iteration, value iteration), temporal difference learning and Q learning. Lecture slides and homework will be released on the Collab.