Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 53 of the 55 references it cites were found in a registry or an index; 49 of the 81 sources its claims cite were checked for support.

Quantum annealing for combinatorial optimization: a benchmarking study

Open access
Kim, Seongmin, Ahn, Sang-Woo, Suh, In-Saeng, Dowling, Alexander W., Lee, Eungkyu, Luo, Tengfei · 2025 · npj Quantum Information

Abstract

Quantum annealing (QA) has the potential to significantly improve solution quality and reduce time complexity in solving combinatorial optimization problems compared to classical optimization methods. However, due to the limited number of qubits and their connectivity, the QA hardware did not show such an advantage over classical methods in past benchmarking studies. Recent advancements in QA with more than 5000 qubits, enhanced qubit connectivity, and the hybrid architecture promise to realize the quantum advantage. Here, we use a quantum annealer with state-of-the-art techniques and benchmark its performance against classical solvers. To compare their performance, we solve over 50 optimization problem instances represented by large and dense Hamiltonian matrices using quantum and classical solvers. The results demonstrate that a state-of-the-art quantum solver has higher accuracy (~0.013%) and a significantly faster problem-solving time (~6561×) than the best classical solver. Our results highlight the advantages of leveraging QA over classical counterparts, particularly in hybrid configurations, for achieving high accuracy and substantially reduced problem solving time in large-scale real-world optimization problems.

Overview

CoChat Summary

## TL;DR

This paper benchmarks quantum annealing (QA), hybrid QA (HQA), and several classical solvers on large, dense QUBO optimization problems meant to reflect real-world materials design tasks. The authors report that HQA was the best overall method, with about 0.013% higher accuracy and about 6561× faster solving time than the best classical solver on 10,000-variable problems.

## Key findings

  • Across large and dense QUBO problems, HQA achieved the highest accuracy, reported as about 0.013% improvement over the best classical solver.
  • For 10,000-dimensional QUBO problems, HQA was about 6561× faster than the best classical solver.
  • Classical solvers (IP, SA, SD, TS, PT-ICM) generally became less accurate and slower as problem size increased; the paper notes that SA and PT-ICM struggled beyond 3000 variables.
  • For one case at n = 7000, the IP solver still had an optimality gap of about 17.73% after 2 hours.
  • Example solving times at n = 5000 were 0.0854 s for HQA, 74.59 s for QA-QBSolv, 167.4 s for SA-QBSolv, and 195.1 s for PT-ICM-QBSolv; at n = 10,000, times were 0.0855 s for HQA, 101 s for QA-QBSolv, and 561 s for SA-QBSolv.

## Methods

  • Benchmarking study comparing seven solvers: QA, HQA, IP, SA, SD, TS, and PT-ICM, with decomposition variants using QBSolv where applicable.
  • Two problem settings were used: real-world-inspired materials optimization QUBOs and systematically generated dense random QUBOs with sizes from 120 to 10,000 variables and four coefficient distributions (standard deviations 0.001, 0.01, 0.1, 1).
  • Main analysis compared relative accuracy and computational time; QA was run on D-Wave Advantage 4.1 with 1000 reads and 20 μs annealing time, while classical solvers used specified default or fixed hyperparameters.

## Limitations

  • The paper states that pure QA and QA with QUBO decomposition still have limitations for large problems because of current hardware limits, especially qubit count and qubit connectivity.
  • The paper notes that the detailed algorithm of HQA is not publicly released.
  • Evident from the text: the benchmarking is limited to QUBO formulations and the specific solver settings/software/hardware described, so the results may not generalize to other optimization formulations or implementations.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s41534-025-01020-1

Cites

53 of 53 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Image 1
From the PDF
Fig. 3 Solving time of the solvers for given QUBO problems. The solving time of ( a ) the classical and quantum solvers and ( b ) the classical solvers (SA, SD, and TS) for small QUBO problems. Quantum solvers do not scale in solving time as the problem size increases, which is a great advantage over classical counterparts.
Fig. 3 Solving time of the solvers for given QUBO problems. The solving time of ( a ) the classical and quantum solvers and ( b ) the classical solvers (SA, SD, and TS) for small QUBO problems. Quantum solvers do not scale in solving time as the problem size increases, which is a great advantage over classical counterparts.From the PDF
Image 3
From the PDF
Fig. 4 Performance of the QA -QBSolv solver with different decomposition sizes. a Relative accuracy and b Solving time of the QA -QBSolv solver for given QUBO problems with different subQUBO sizes.
Fig. 4 Performance of the QA -QBSolv solver with different decomposition sizes. a Relative accuracy and b Solving time of the QA -QBSolv solver for given QUBO problems with different subQUBO sizes.From the PDF

Verification

Claims the paper makes

24 Valid · 5 Not valid · 8 Not checked · 54 Own results and interpretations

30 more claims · 3 with their cited sources checked

Create a free account to see all 91 claims and to verify the 1 reference still open.

See all claims — free account
  • Probably not a substantial claim. 32
  • It may state two separate claims as one. 33

Create a free account to see these claims.

Citations

Cites

Show all 55 references

Research works printed but not linked: 3

  1. Tao, M. et al. Towards Stability in the Chapel Language. In Proc. IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW) 557–566 (IEEE, 2020).
  2. Hansen, P. B. Simulated annealing. Electrical Eng. Comput. Sci. Tech. Rep. 170, 1–12 (1992).
  3. Dupin, N., Nielsen, F. & Talbi, E. Dynamic Programming heuristic for k-means Clustering among a 2-dimensional Pareto Frontier. In Proc. 7th International Conference on Metaheuristics and Nature Inspired Computing, Springer (2018).

Cited by

See all 46

Metadata

Authors
Kim, S., Ahn, S.-W., Suh, I.-S., Dowling, A. W., Lee, E., & Luo, T.
Affiliations
  • University of Notre Dame (US)
  • Kyung Hee University (KR)
  • Oak Ridge National Laboratory (US)
Corresponding author
  • Dowling, Alexander W.
  • Lee, Eungkyu
  • Luo, Tengfei
Venue
npj Quantum Information
Published
2025-05-16
Volume
11
Issue
1
ISSN
2056-6387 (ISSN-L)
OpenAlex
W4410453046
Cited by
46
References
55
Citations by year
2026: 35 · 2025: 11
FWCI
18.31
Type
journal article
Language
en
Open access
Licence
Locations
  • npj Quantum Information (publishedVersion, cc-by-nc-nd)
  • arXiv (Cornell University) (submittedVersion, cc-by-nc-nd)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • OSTI OAI (U.S. Department of Energy Office of Scientific and Technical Information) (submittedVersion)
  • arXiv (Cornell University)
Funders
  • U.S. Department of Energy (DE-AC05-00OR22725)
  • National Research Foundation of Korea (RS-2023-00255442)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Quantum Computing Algorithms and Architecture
  • Physical Sciences › Computer Science › Artificial Intelligence › Metaheuristic Optimization Algorithms Research
  • Physical Sciences › Mathematics › Numerical Analysis › Advanced Optimization Algorithms Research
Keywords
quantum annealing, combinatorial optimization, benchmarking, quantum advantage, hybrid quantum-classical solvers, qubit connectivity, solution accuracy, problem-solving time, large-scale optimization problems

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-09-30, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-09-30 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Quantum annealing for combinatorial optimization: a benchmarking study. https://app.cochat.ai/papers/10.1038/s41534-025-01020-1
The paper
Kim, S., Ahn, S.-W., Suh, I.-S., Dowling, A. W., Lee, E., & Luo, T. (2025). Quantum annealing for combinatorial optimization: a benchmarking study. npj Quantum Information. https://doi.org/10.1038/s41534-025-01020-1

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.