Management Sciences, University of Waterloo
Date: Friday, February 10, 2017 at 1 pm.
Location: DC 1331
Title: Semidefinite Optimization for Energy Networks
Abstract: In many energy network optimization problems, the decision maker faces the combined challenge of nonlinear constraints, discrete decisions, system parameters affected by uncertainty, and the scale of the underlying network. However, such problems also exhibit structure which can be exploited in convex relaxations in order to improve the solution quality and the scalability of the optimization solvers. In this work, we focus on the AC optimal power flow problem which appears as the backbone of power systems operations, hence, solving this problem to optimality is critical for energy applications. We present semidefinite-based relaxations to solve the optimal power flow problem and its variants including transmission network expansion planning and unit-commitment with AC optimal power flow constraints. Computational experiments show that the proposed formulations and algorithms efficiently solve IEEE instances and lead to significant cost benefits with provably tight bounds.
Bio: Bissan Ghaddar is an Assistant Professor in Data Analytics and Operations Research at the University of Waterloo working on problems at the intersection of smart cities, IoT, data analytics, and mathematical modeling and optimization for large-scale problems. Prior to joining the University of Waterloo, she worked on energy and transportation network optimization at IBM Research Dublin and on inventory management problems at the Centre for Operational Research and Analysis, Department of National Defence Canada. She was also invited for extended research visits at the Universität zu Köln in Germany and the University of Avignon in France. Dr. Ghaddar received a Ph.D. degree in operations research from the University of Waterloo, Canada. Her research has been supported by prestigious national and international scholarships and in 2012 she was awarded an FP7 IIF European Union Grant.