3.0

credits

Average Course Rating

The goal of this course is to explore cutting-edge methods in power system optimization and machine learning, aimed at solving industry-grade decision support challenges related to optimal power flow and unit commitment. The course begins by building foundational knowledge of the mathematical formulations underpinning these problems. It then delves into optimization methods for their efficient solution with exact, nearly exact, occasionally exact, pt approximate relaxation/restriction techniques. These methods are then applied within the framework of security-constrained optimal power flow, which in turn is then enhanced to deal with parametric uncertainty and unit commitment decisions. The course concludes with an in-depth review of how machine learning techniques can accelerate power flow optimization, focusing on approaches such as learning to optimize and learning to sample, and their benchmarking with conventional optimization approaches in terms of interpretability, trustworthiness, and guarantees.