Date: November 20th, 2026

Time: 3-5pm ET

In-Person Location: George Tower 607

Teams Link: https://teams.microsoft.com/meet/2335662066017656?p=9s6vPpDvo4UH3mUwIb

Rosemary Barrass

Machine Learning PhD Student

H. Milton Stewart School of Industrial and Systems Engineering

Georgia Institute of Technology

Committee

1. Dr. Pascal Van Hentenryck (Advisor), Georgia Institute of Technology

2. Dr. Valerie Thomas, Georgia Institute of Technology

3. Dr. Daniel Molzahn, Georgia Institute of Technology

4. Dr. Yao Xie, Georgia Institute of Technology

5. Dr. José Tabarez, Los Alamos National Laboratory

Abstract

The rapid adoption of renewable generation, data centers, and electric vehicle charging across the U.S. has led to a major shift in grid operations. What was once a predictable, simple grid has turned into a highly volatile, massive system. Every day, grid operators have to schedule generation units using Unit Commitment decisions across the U.S. with limited knowledge of the exact grid conditions ahead. Researchers have spent decades studying this Unit Commitment problem under uncertainty to reduce the computational complexity, improve solution times, and more accurately model real grid conditions. Despite these efforts, lack of open-source data continues to hinder research on the most complex generators in the U.S.: combined-cycle generators. This work seeks to fill this gap with novel techniques that decompose and speed-up the solution of Unit Commitment problems with combined-cycle generators while simultaneously looking ahead to new technologies that can help shape the future of power systems optimization: quantum computing. By harnessing both quantum computing and advanced decomposition methods, this work demonstrates effective new tools to combat the twin scheduling problems of high volatility and massive grid scale.