A Fast Quantum Algorithm for Searching the Quasi-Optimal Solutions of Unit Commitment

The electric power industry faces tough optimization problems that are challenging for classical computers, e.g., the unit commitment (UC). Quantum computing has the potential to provide a speedup for combinatorial optimization problems utilizing ad hoc algorithms like Grover's search algorithm. The quantum oracle is indispensable for Grover's search, yet it is difficult to realize. In this letter, we develop a way to construct efficient oracles for solving UC problems in Grover's framework. The oracle features a physics-informed circuit and a variational quantum circuit picking up good solutions. A quantum algorithm that leverages the oracle is then proposed, which could reduce the number of tractable linear programming subproblems required to solve the targeted UC problem when working in conjugation with classical computers. The oracle and algorithm are validated by results obtained on real quantum hardware and simulators.

I. Introduction

Unit commitment (UC) is a fundamental optimization problem in power systems that is solved multiple times daily to obtain a cost-effective schedule of generators. UC is typically formulated as a constrained mixed-integer linear program (MILP) and solved with the Branch and Cut (B&C) method on general-purpose solvers. Even though there is a practical demand on improving the model accuracy and solution optimality of UC [1], the capacity limit of MILP solvers hinders system operators from making such improvements.