Ising formulation of integer optimization problems for utilizing quantum annealing in iterative improvement strategy

TL;DR

This paper introduces an Ising formulation for integer optimization problems that leverages quantum annealing's biased sampling to improve solutions iteratively, enabling hybrid approaches with classical algorithms.

Contribution

The authors propose a novel Ising formulation that allows quantum annealing to be used in iterative improvement strategies for integer optimization problems.

Findings

Biased sampling of degenerated ground states is exploited in the formulation.

Avoidance of first-order phase transition in a fully connected ferromagnetic Potts model.

The formulation is applicable to a wide range of integer optimization problems.

Abstract

Quantum annealing is a heuristic algorithm for searching the ground state of an Ising model. Heuristic algorithms aim to obtain near-optimal solutions with a reasonable computation time. Accordingly, many algorithms have so far been proposed. In general, the performance of heuristic algorithms strongly depends on the instance of the combinatorial optimization problem to be solved because they escape the local minima in different ways. Therefore, combining several algorithms to exploit their complementary strength is effective for obtaining highly accurate solutions for a wide range of combinatorial optimization problems. However, quantum annealing cannot be used to improve a candidate solution obtained by other algorithms because it starts from an initial state where all spin configurations are found with a uniform probability. In this study, we propose an Ising formulation of integer optimization problems to utilize quantum annealing in the iterative improvement strategy. Our formulation exploits the biased sampling of degenerated ground states in transverse magnetic field quantum annealing. We also analytically show that a first-order phase transition is successfully avoided for a fully connected ferromagnetic Potts model if the overlap between a ground state and a candidate solution exceeds a threshold. The proposed formulation is applicable to a wide range of integer optimization problems and enables us to hybridize quantum annealing with other optimization algorithms.