AEEM 9074: Reinforcement Learning and Data-driven Control

This course introduces core ideas in data-driven control and reinforcement learning for dynamical systems. We begin with the motivation for designing controllers directly from measured input and output data, then study foundational concepts such as persistency of excitation and Willems' fundamental lemma. These tools provide a bridge between experimental trajectories and system-theoretic representations that can be used for prediction, analysis, and controller synthesis. We then discuss reinforcement learning methods for control, including how value functions, policies, exploration, and data-driven optimization can be used to improve closed-loop performance when accurate models are unavailable or difficult to obtain.