Transfer learning of optimal QAOA parameters in combinatorial optimization
The DOI matches its Crossref record; 38 of the 38 references it cites were found in a registry or an index.
Abstract
Solving combinatorial optimization problems (COPs) is a promising application of quantum computation, with the quantum approximate optimization algorithm (QAOA) being one of the most studied quantum algorithms for solving them. However, multiple factors make the parameter search of the QAOA a hard optimization problem. In this work, we study transfer learning (TL), a methodology to reuse pre-trained QAOA parameters of one problem instance into different COP instances. This methodology can be used to alleviate the necessity of classical optimization to find good parameters for individual problems. To this end, we select small cases of the traveling salesman problem (TSP), the bin packing problem (BPP), the knapsack problem (KP), the weighted maximum cut (MaxCut) problem, the maximal independent set (MIS) problem, and portfolio optimization (PO), and find optimal β\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\beta $$\end{document} and γ\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\gamma $$\end{document} parameters for p layers. We compare how well the parameters found for one problem adapt to the others. Among the different problems, BPP is the one that produces the best transferable parameters, maintaining the probability of finding the optimal solution above a quadratic speedup over random guessing for problem sizes up to 42 qubits and p=10\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$p = 10$$\end{document} layers. Using the BPP parameters, we perform experiments on IonQ Harmony and Aria, Rigetti Aspen-M-3, and IBM Brisbane of MIS instances for up to 18 qubits. The results indicate that IonQ Aria yields the best overlap with the ideal probability distribution. Additionally, we show that cross-platform TL is possible using the D-Wave Advantage quantum annealer with the parameters found for BPP. We show an improvement in performance compared to the default protocols for MIS with up to 170 qubits. Our results suggest that there are QAOA parameters that generalize well for different COPs and annealing protocols.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“While this mitigation technique can correct errors coming from the readout of the quantum device,”
From Toward a linear-ramp QAOA protocol: evidence of a scaling advantage in solving some combinatorial optimization problems (2025) · cites this paper as [20]
Citations
Cites
Variational quantum algorithms
Cerezo, M., Arrasmith, A., Babbush, R., Benjamin, S. C., Endo, S., Fujii, K., McClean, J. R., Mitarai, K., Yuan, X., Cincio, L., et al. (2021) · Nature Reviews Physics · doi:10.1038/s42254-021-00348-9 · Cited by 3,385 · cited 1×
“The underlying idea of parameter transfer is very promising, as it has been shown that finding parameters for variation quantum algorithms, such as QAOA, is NP-hard [10].”
Ising formulations of many NP problems
Lucas, A. (2014) · Frontiers in Physics · doi:10.3389/fphy.2014.00005 · Cited by 2,512 · cited 2×
“The TSP formulation is given by min x i,t = 1 ∀ t = 0 , ..., n − 1 , (A10)”
Quantum annealing with manufactured spins
Johnson, M. W., Amin, M. H. S., Gildert, S., Lanting, T., Hamze, F., Dickson, N., Harris, R., Berkley, A. J., Johansson, J., Bunyk, P., et al. (2011) · Nature · doi:10.1038/nature10012 · Cited by 1,891 · cited 1×
“The source code for the results shown here can be found at https://jugit.fz-juelich.de/qip/ transfer-learning-QAOA .”
Zhou, L., Wang, S.-T., Choi, S., Pichler, H., & Lukin, M. D. (2020) · Physical Review X · doi:10.1103/physrevx.10.021067 · Cited by 913 · cited 2×
“In [14], numerical evidence supporting an approximation ratio exceeding the worst-case scenario by Goemans-Williamson (GW) was presented for parameters transferred in 3-regular graphs of MaxCut, specifically for p 12.”
Quantum approximate optimization of non-planar graph problems on a planar superconducting processor
Harrigan, M. P., Sung, K. J., Neeley, M., Satzinger, K. J., Arute, F., Arya, K., Atalaya, J., Bardin, J. C., Barends, R., Boixo, S., et al. (2021) · Nature Physics · doi:10.1038/s41567-020-01105-y · Cited by 564 · cited 1×
“Third, quantum algorithms to solve these problems need few resources and can be tested on current state-of-the-art quantum hardware [4–6].”
The unconstrained binary quadratic programming problem: a survey
Kochenberger, G., Hao, J.-K., Glover, F., Lewis, M., Lü, Z., Wang, H., & Wang, Y. (2014) · Journal of Combinatorial Optimization · doi:10.1007/s10878-014-9734-0 · Cited by 473
Training Variational Quantum Algorithms Is NP-Hard
Bittel, L., & Kliesch, M. (2021) · Physical Review Letters · doi:10.1103/physrevlett.127.120502 · Cited by 398 · cited 2×
“Subse-quently, [12] extended this concept to non-isomorphic unweighted graphs in the MaxCut problem and showed optimal QAOA parameter concentration for large graphs.”
JUWELS: Modular Tier-0/1 Supercomputer at the Jülich Supercomputing Centre
Krause, D. (2019) · Journal of large-scale research facilities JLSRF · doi:10.17815/jlsrf-5-171 · Cited by 271 · cited 1×
“In Appendix A, we explain in detail the problems and the parameters used in this work.”
Warm-starting quantum optimization
Egger, D. J., Mareček, J., & Woerner, S. (2021) · Quantum · doi:10.22331/q-2021-06-17-479 · Cited by 269 · cited 1×
“To some extent, QAOA can be seen as a trotterized quantum annealing protocol where the number of layers p determines the precision of the solution [8].”
Quantum critical dynamics in a 5,000-qubit programmable spin glass
King, A. D., Raymond, J., Lanting, T., Harris, R., Zucca, A., Altomare, F., Berkley, A. J., Boothby, K., Ejtemaee, S., Enderud, C., et al. (2023) · Nature · doi:10.1038/s41586-023-05867-2 · Cited by 242 · cited 1×
“The source code for the results shown here can be found at https://jugit.fz-juelich.de/qip/ transfer-learning-QAOA .”
Research works printed but not linked: 6
- Farhi, E., Goldstone, J., Gutmann, S.: A quantum approximate optimization algorithm (2014), arXiv:1411.4028 [quant-ph] · arXiv
- Koßmann, G., Binkowski, L., van Luijk, L., Ziegler, T., Schwonnek, R.: Deep-circuit qaoa (2023). arXiv:2210.12406 [quant-ph] · arXiv
- Boulebnane, S., Montanaro, A.: Solving boolean satisfiability problems with the quantum approximate optimization algorithm (2022), arXiv:2208.06909 [quant-ph] · arXiv
- Wurtz, J., Lykov, D.: The fixed angle conjecture for qaoa on regular maxcut graphs (2021), arXiv:2107.00677 [quant-ph] · arXiv
- Powell, M.: A direct search optimization method that models the objective and constraint functions by linear interpolation. Adv. Optim. Numer. Anal. 275, 51 (1994)
- Cheeseman, P., Kanefsky, B., Taylor, W. M.: Where the Really Hard Problems Are, The 12nd International Joint Conference on Artificial Intelligence , 331 (1991)
Cited by
Montañez-Barrera, J. A., & Michielsen, K. (2025) · npj Quantum Information · doi:10.1038/s41534-025-01082-1 · Cited by 34
“Usually, H C is derived from the quadratic unconstrained binary optimization (QUBO) formulation 2,20,39 .”
Self-citationQuantum approximate multi-objective optimization
Kotil, A., Pelofske, E., Riedmüller, S., Egger, D. J., Eidenbenz, S., Koch, T., & Woerner, S. (2025) · Nature Computational Science · doi:10.1038/s43588-025-00873-y · Cited by 63
“Another promising strategy is to reuse optimized parameters from similar problem instances [35–37].”
Andreou, A., Mavromoustakis, C. X., Mastorakis, G. N., Bourdena, A., & Markakis, E. (2025) · IEEE Access · doi:10.1109/access.2025.3603338 · Cited by 18
“QAOA constructs a parameterized quantum circuit that alternates between applying the cost Hamiltonian and a mixing Hamiltonian [35].”
Achamrah, F. E. (2024) · IEEE Access · doi:10.1109/access.2024.3505678 · Cited by 10
Towards Robust Benchmarking of Quantum Optimization Algorithms
Bucher, D., Kraus, N., Blenninger, J., Lachner, M., Stein, J., & Linnhoff–Popien, C. (2024) · IEEE International Conference on Quantum Computing and Engineering (QCE) · doi:10.1109/qce60285.2024.11030870 · Cited by 16
“However, parameter optimization can also be seen as a generalization task similar to machine learning [65].”
Metadata
- Authors
- Montañez-Barrera, J. A., Willsch, D., & Michielsen, K.
- Affiliations
- Forschungszentrum Jülich (DE)
- Corresponding author
- Montañez-Barrera, J. A.
- Venue
- Quantum Information Processing
- Published
- 2025-05-09
- Volume
- 24
- Issue
- 5
- ISSN
- 1570-0755 (ISSN-L)
- 1573-1332
- OpenAlex
- W4410232493
- Cited by
- 19
- References
- 35
- Citations by year
- 2026: 8 · 2025: 8
- FWCI
- 6.42
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- www.springernature.com/gp/researchers/text-and-data-mining (tdm)
- www.springernature.com/gp/researchers/text-and-data-mining (vor)
- Locations
- Quantum Information Processing (publishedVersion)
- Funders
- Bundesministerium für Bildung und Forschung (13N16149)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Quantum Computing Algorithms and Architecture
- Physical Sciences › Computer Science › Artificial Intelligence › Metaheuristic Optimization Algorithms Research
- Physical Sciences › Physics and Astronomy › Atomic and Molecular Physics, and Optics › Quantum many-body systems
- Keywords
- transfer learning, quantum approximate optimization algorithm, combinatorial optimization, bin packing, maximal independent sets, traveling salesman problem, knapsack problem, portfolio optimization, quantum annealing, D-Wave Advantage, IonQ, IBM
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.
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: Transfer learning of optimal QAOA parameters in combinatorial optimization. https://app.cochat.ai/papers/10.1007/s11128-025-04743-4
- The paper
- Montañez-Barrera, J. A., Willsch, D., & Michielsen, K. (2025). Transfer learning of optimal QAOA parameters in combinatorial optimization. Quantum Information Processing. https://doi.org/10.1007/s11128-025-04743-4
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.