CPS: Small: Infusing Quantum Computing, Decomposition, and Learning for Addressing Cyber-Physical Systems Optimization Challenges

Louisiana State University, Baton Rouge LA

Investigators

Abstract

This NSF project aims to develop quantum computing-inspired algorithms for optimization problems that appear in various critical infrastructure systems, including power systems. The primary focus will be on computationally expensive mixed-integer optimization problems, which are hard to solve by classical computing machines. The project will bring transformative change in the ways that electric power system optimization problems are solved. This will be achieved by integrating machine learning and distributed computing with quantum computing techniques. The intellectual merits of the project include advancing the understanding of quantum computing optimization algorithms and harnessing their potential for solving complex problems. The broader impacts of the project include facilitating quantum computing applications for solving a variety of optimization problems, particularly for power systems operation and planning. In addition, multiple educational and outreach activities are envisioned to broaden the participation of under-resourced and underrepresented groups in computing and engineering, contribute to diverse workforce development, raise awareness of quantum technology among diverse audiences and children, and benefit local communities and the nation. This project will contribute to the field of cyber-physical systems optimization. Several existing technical and algorithmic challenges for solving optimization problems with quantum computing will be addressed. The project proposes the development of decomposition and distributed approaches, combined with quantum computing and machine learning, to create trainable quantum-classical algorithms for various mixed-binary optimization problems. The proposed strategies will enable the conversion of deterministic and stochastic mixed-binary problems in the power grid domain into multiple subproblems. These subproblems will be solvable by quantum computers in collaboration with classical computing machines. To achieve this, trainable variational quantum algorithms will be developed and integrated into intelligent quantum Lagrangian and Benders techniques. This integration will coordinate the computing operations performed by both quantum and classical machines. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.