Pith. sign in

REVIEW 3 major objections 4 minor 72 references

Variational matrix product states for combinatorial optimization

T0 review · 3 major / 4 minor · reviewed 2026-08-03 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper argues that variational matrix product state methods, when wrapped in iterated local search, outperform both classical heuristics like ILS and quantum algorithms like QAOA on large MaxCut problems.

desk verdict A clearly-constructed tensor-network heuristic whose headline Gset advantage is not yet trustworthy because the comparison is per-instance tuned and the reported hyperparameters are inconsistent. read the letter →

arxiv 2512.20613 v2 pith:RGOCJOKK submitted 2025-12-23 quant-ph physics.comp-ph

Guillermo Preisser , Conor Mc Keever , Michael Lubasch This is my paper · ORCID
classification quant-phphysics.comp-ph
keywords MaxCutmatrixproductstatesiteratedlocalsearchquantum-inspiredoptimizationvariationalmethodscombinatorialGPUparallelizationQAOA
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The authors show that combining variational product-state or matrix-product-state minimization of a quantum-annealing Hamiltonian with iterated local search yields a solver, QiILS, that finds better MaxCut solutions than standard ILS, LQA, GCS, and QAOA on instances up to 50,000 variables. A parallel variant, QiIGS, achieves comparable accuracy while running an order of magnitude faster on a 20,000-variable graph. A sympathetic reader would care because the result suggests that classical, quantum-inspired tensor-network heuristics remain strong competitors to near-term quantum optimization algorithms.

What carries the argument

The central object is the interpolating Hamiltonian H(λ) = (1-λ)Hi + λHf, where Hi is a transverse-field Hamiltonian and Hf encodes the MaxCut objective. The QiILS algorithm alternates between variational energy minimization of this Hamiltonian at a fixed λ and random spin flips (perturbations), with the product-state variant updating each angle θj via a closed-form formula. QiIGS replaces those sequential updates with parallel gradient descent, enabling GPU acceleration.

What would settle it

Take a collection of unseen Gset instances, choose λ once on a separate training set (e.g., via golden-section search), then run QiILS and ILS with that fixed λ and charge all tuning time to QiILS. If QiILS no longer achieves consistently lower relative error than ILS, the paper's main advantage claim is falsified.

Watch

Extended reading notes

Core claim

The paper's central claim is that optimizing a quantum annealing interpolation Hamiltonian with DMRG-style sweeps, followed by random spin perturbations, produces a strictly generalized form of ILS that systematically improves approximation ratios on MaxCut. The authors report that the unentangled product-state version (bond dimension 1) with many iterations outperforms higher bond dimensions on large instances, and that QiILS beats ILS, LQA, GCS, and QAOA on Gset benchmarks. They further claim that QiIGS, which replaces sequential angle updates with parallel gradient descent, scales nearly independently of problem size on GPU hardware and provides an order-of-magnitude speedup over QiILS on

Load-bearing premise

The load-bearing premise is that per-instance hyperparameter tuning (especially λ) is a fair part of the method's cost and that the tuned values generalize; if a held-out evaluation with fixed hyperparameters erases the gap over ILS, the central performance claim collapses.

Editorial extensions

If this is right

  • QiILS recovers standard ILS at λ=1, so any improvement must come from the annealing-like intermediate states sampled at λ<1.
  • On unweighted 3-regular graphs, QiILS with χ=1 solves all 1000 tested instances within 15 iterations, a performance the authors compare favorably to QAOA's required circuit depth.
  • Per-sweep wall-clock time of QiILS matches that of ILS, while LQA is roughly 7× slower and GCS roughly 9000× slower on the G12 benchmark.
  • QiIGS's parallel gradient updates allow near-constant per-iteration time as the problem grows to 50,000 variables.
  • The same recipe—annealing interpolation plus iterated local search—could be applied to other binary optimization problems beyond MaxCut.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The reported advantage of QiILS over ILS depends on per-instance tuning of λ (and perturbation strength p); if tuning overhead is charged to QiILS or λ is fixed by a held-out selection, the margins in Table I and Fig. 1 could shrink.
  • The strong performance of the product-state (χ=1) version suggests that entanglement captured by larger bond dimensions is not the driver of the gains; the annealing path plus ILS perturbation loop may carry most of the benefit.
  • These quantum-inspired solvers could serve as a stronger classical baseline for future demonstrations of quantum optimization advantage, since they already outperform a popular variational quantum algorithm on tested instances.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper introduces quantum-inspired variational algorithms for MaxCut based on product-state (PS) and matrix-product-state (MPS) ansätze, embedded in iterated local search (ILS). The single-site PS update is derived exactly (Eqs. 4–6), and MPS versions are implemented with DMRG. The authors also introduce QiIGS, a parallel gradient-descent variant, and benchmark these methods on random regular graphs, Gset instances up to 800 vertices, and one 20,000-vertex instance (G81). The paper claims that QiILS outperforms classical ILS, local quantum annealing (LQA), generalized coherent states (GCS), and QAOA on the tested instances, and that QiIGS offers an order-of-magnitude speedup over QiILS on G81.

Significance. If the empirical claims are robust, the paper would make a useful contribution: the exact PS updates are simple, the algorithms are classically scalable, and the idea of combining quantum-inspired variational evolution with ILS is natural and interesting. The derivation is transparent, and the numerical study is extensive, including hyperparameter exploration and a large-scale GPU comparison. However, the reported advantage over baselines rests on a comparison in which QiILS's key hyperparameter λ is tuned per target instance, while the baselines use fixed hyperparameters. This, together with internal inconsistencies in the reported hyperparameters and an overstated characterization of the baselines, substantially weakens the central claim as currently presented.

major comments (3)
  1. [Supplemental Sec. IV, Table S2; Table I; Fig. 2(a)] The Gset comparison is not on equal footing: per-instance λ values for QiILS (e.g., G1=0.38, G2=0.41, G6=0.42) were selected using 10 random initializations and 50 iterations on the same graph being evaluated (Supplemental Sec. IV), while LQA and GCS use fixed hyperparameters. Since Fig. 2(a) shows that performance varies by orders of magnitude with λ, the advantage in Table I and Fig. 1 may reflect per-instance selection rather than algorithmic superiority. Please provide a train/validation split, held-out instances, or explicitly charge the tuning cost to QiILS; without this, the main empirical claim is not established.
  2. [Fig. 3(b) vs Table S2; Fig. 2(b) vs Table S2; Table I vs Table S2] The reported hyperparameters are internally inconsistent. The Fig. 3(b) caption states λ=0.75, p=0.3, while Table S2 lists λ=0.65, p=0.5; the Fig. 2(b) caption says λ=0.4, while Table S2 says λ=0.5; and Table I says QiILS uses sweeps=80, while Table S2 lists sweeps=200 for G1–G10. These discrepancies make the protocol non-reproducible and must be reconciled before the results can be used to support the conclusions.
  3. [Conclusions] The conclusions call ILS, LQA, and GCS 'state-of-the-art classical heuristics.' This is overstated: ILS is a generic metaheuristic, LQA and GCS are quantum-inspired variational methods, and no leading classical MaxCut solvers (e.g., Breakout Local Search, simulated annealing, or commercial solvers) are compared. The competitive claim should be scoped to the tested baselines, not to the state of the art in combinatorial optimization.
minor comments (4)
  1. [Background and Figs. 2, 5] The abbreviations u3R and w3R are used in figures but not defined; the text defines udR and wdR. Use consistent notation.
  2. [Eq. (8)] The rounding rule maps θ_j to bits, but the relation between the bit b_j and the spin value (±1) used in the Hamiltonian is not explicitly stated; clarify to avoid ambiguity.
  3. [Supplemental Sec. IV] The Supplemental Material is referenced only by a '[URL will be inserted by publisher]' placeholder; a working link or an included supplemental PDF is needed for reproducibility.
  4. [Table S2] The table's 'iterations' column is not consistently defined for all rows; for example, Fig. 2(a) lists 1,2,4,8 while Fig. 4 lists 10,000. Clarify whether these are iteration counts or iteration indices.

Circularity Check

1 steps flagged · score 2.0 of 10

Core variational derivation is self-contained; the main circularity-adjacent issue is per-instance λ tuning in the Gset benchmark, not the algorithm construction.

  1. fitted input called prediction [Supplemental Sec. IV; Table I caption; End Matter 'Additional Gset results']
    "Fixed hyperparameters: for QiILS, sweeps = 80, p = 0.3; ... To find the optimal λ values for the Gset calculations, 10 random initializations and 50 iterations were used for each graph. ... We conclude from Tab. I that QiILS consistently outperforms the other solvers, in line with our conclusions drawn from Fig. 1."

    QiILS's key hyperparameter λ is selected for each Gset graph by running QiILS on that same graph (Supplemental Sec. IV: '10 random initializations and 50 iterations were used for each graph'), and Table S2 lists a different λ per graph. The reported conclusion that QiILS 'consistently outperforms' ILS/LQA/GCS is then drawn from these per-instance fitted λ values, while the baselines are run with fixed hyperparameters. Thus the headline empirical comparison is a per-instance best-hyperparameter curve against fixed-parameter competitors: a fitted input is presented as an independent predictive advantage. This does not infect the Eq. (4)–(6) derivation, but it makes the benchmark claim partly self-selected.

full rationale

The algorithmic derivation is self-contained. Equation (4) is the exact energy expectation value of H(λ) in the product-state ansatz, Eq. (5) defines A from the graph weights and other angles, and Eq. (6) is the exact minimizer of Eq. (4). The rounding rule (8) and the σX perturbation are explicit operational steps, and QiIGS's Eqs. (9)–(10) are gradient descent on the same energy. No step in this chain invokes the benchmark conclusions, and no load-bearing self-citation is used; the cited prior tensor-network, QAOA, LQA, and GCS works are external. The only circularity-adjacent feature is in the benchmarking protocol: λ (and to some extent p and sweeps) are tuned per Gset instance on the same instances that are later reported, while Table I's caption reports fixed hyperparameters and Table S2 contradicts it (e.g., sweeps 80 vs 200; Fig. 3(b) λ = 0.75 vs Table S2 λ = 0.65). This is a fairness/reproducibility issue for the empirical performance claim rather than a derivation-level circularity, so the score remains low.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central algorithm is variational optimization of PS/MPS ansatze on a fixed-lambda annealing Hamiltonian, wrapped in ILS. The ledger shows that the main free knobs are lambda, p, sweeps, tau, and convergence tolerance; none of these are derived from first principles, and all are tuned on the benchmark instances. No new physical entities are postulated.

free parameters (6)
  • Annealing interpolation lambda = Per-instance/per-figure values; e.g. G1=0.38, G2=0.41, G6=0.42, G81=0.35; Fig. 3 uses ~0.55 (u3R) and ~0.65/0.75 (w3R, i
    Selected by golden-section search/grid to maximize the fitted decay rate of mean energy; for Gset it is tuned per instance on the same instance then evaluated, so the benchmark includes test-instance information.
  • Perturbation strength p = 0.1–0.7 depending on figure and graph type; e.g. p=0.5 (Fig. 2/3 u3R), p=0.3 (w3R), p=0.15 (Fig. 4)
    Chosen by hyperparameter search; the paper shows performance is highly sensitive to p (Fig. 2c), making it a load-bearing tuning knob.
  • Max sweeps per iteration = 80 or 200
    User-set cap on DMRG/PS sweeps; affects quality/runtime trade-off and is not derived from any principle.
  • QiIGS step size tau = 0.1 (Fig. 4)
    Chosen by hand for the large-scale GPU runs; no schedule or adaptation is described.
  • Convergence tolerance epsilon in Eq. (7) = Not specified
    User-defined tolerance in the PS convergence criterion; unstated value prevents exact reproduction.
  • Decay-rate fit constants c0, c1 in lambda selection = Fitted per lambda candidate
    Used in the exponential fit E ~ c0 e^{-m*iota} + c1 to choose lambda; these are fitting parameters, not derived quantities.
assumptions (4)
  • domain assumption DMRG with ITensor converges to a good approximation of the ground state of H(lambda) for the graphs considered.
    The entire sampling/perturbation loop relies on MPS optimizations finding low-energy states at fixed lambda; convergence is observed, not proven.
  • domain assumption The greedy bitstring extraction from the variational state (MPS sequential sampling or PS rounding Eq. (8)) yields cut values whose quality tracks the variational energy.
    Used without proof; a low-energy state could in principle project onto a poor cut if sampling/rounding is biased.
  • ad hoc to paper The exponential decay model E ~ c0 e^{-m*iota} + c1 used for lambda selection is a valid proxy for solver performance.
    Introduced specifically to justify lambda selection; no theoretical reason is given for exponential decay, and the selected lambda depends on this fitting model.
  • standard math Standard spin mapping C = (sum_{j<k} w_{j,k} - E)/2 and single-site PS minimization (Eqs. 4-6) are valid.
    Unproved background used as input; these are standard identities for Ising MaxCut encodings and product-state energy minimization.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Variational matrix product states for combinatorial optimization." pith.science (2026). https://pith.science/paper/RGOCJOKK

@misc{pith2026251220613,
  author       = {Pith},
  title        = {Pith review of: Variational matrix product states for combinatorial optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RGOCJOKK}},
  note         = {Machine review of arXiv:2512.20613}
}
read the original abstract

To compute approximate solutions for combinatorial optimization problems, we describe variational methods based on the product state (PS) and matrix product state (MPS) ans\"{a}tze. We perform variational energy minimization with respect to a quantum annealing Hamiltonian and utilize randomness by embedding the approaches in the metaheuristic iterated local search (ILS). The resulting quantum-inspired ILS algorithms are benchmarked on maximum cut problems of up to 50000 variables. We show that they can outperform traditional (M)PS methods, classical ILS, the quantum approximate optimization algorithm and other variational quantum-inspired solvers.

Figures

Figures reproduced from arXiv: 2512.20613 by the authors.

Figure 1
Figure 1. FIG. 1. Performance comparison for G12. Average rela [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Study of QiILS hyperparameters. Performance is [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Comparison with QAOA. Performance is measured [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: FIG. 4. Comparison of QiIGS with QiILS. In the main [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

72 extracted references · 10 linked inside Pith

  1. [1]

    QiIGS —The QiIGS algorithm replaces the sequential QiILS update of Eq

    The detailed analysis and all results are provided in the End Matter. QiIGS —The QiIGS algorithm replaces the sequential QiILS update of Eq. ( 6) by a global update based on gradient descent: θnew j =θj −τ∂E ∂θj , (9) where τ denotes the step size, ∂E ∂θj = −2λa sin(2θj) − 2(1 −λ) cos(2θj) (10) (a) (b) 𝜄 0 5 10 15 20 1 − r 10 0 10 − 1 10 − 2 10 − 3 10 − 4...

    2023

  2. [2]

    Starting from a randomly initialized MPS with bond dimension χ, we perform DMRG sweeps at the chosen λ until conver- gence

    at a fixed value of λ, we perform an iterative procedure indexed by ι, where each iteration consists of a DMRG optimization followed by a perturbation step, in analogy with ILS. Starting from a randomly initialized MPS with bond dimension χ, we perform DMRG sweeps at the chosen λ until conver- gence. Once converged, we generate a spin configuration (bitstri...

  3. [3]

    Zhang, Y

    C. Zhang, Y. Wu, Y. Ma, W. Song, Z. Le, Z. Cao, and J. Zhang, A review on learning to solve combinatorial op- timisation problems in manufacturing, IET Collab. Intell. Manuf. 5, e12072 (2023)

    2023

  4. [4]

    Guihaire and J.-K

    V. Guihaire and J.-K. Hao, Transit network design and scheduling: A global review, Transp. Res. A 42, 1251 (2008)

    2008

  5. [5]

    R. Z. Farahani, E. Miandoabchi, W. Y. Szeto, and H. Rashidi, A review of urban transportation network design problems, Eur. J. Oper. Res. 229, 281 (2013)

    2013

  6. [6]

    Liu, T.-H

    Y.-F. Liu, T.-H. Chang, M. Hong, Z. Wu, A. Man- Cho So, E. A. Jorswieck, and W. Yu, A Survey of Recent Advances in Optimization Methods for Wireless Commu- nications, IEEE J. Sel. Areas Commun. 42, 2992 (2024)

    2024

  7. [7]

    H. R. Lourenço, O. C. Martin, and T. Stützle, Iterated Local Search: Framework and Applications, in Hand- book of Metaheuristics , edited by M. Gendreau and J.-Y. Potvin (Springer International Publishing, Cham, 2019) pp. 129–168

    2019

  8. [8]

    Kadowaki and H

    T. Kadowaki and H. Nishimori, Quantum annealing in the transverse Ising model, Phys. Rev. E 58, 5355 (1998)

    1998

Show all 72 references
  1. [9]

    Farhi, J

    E. Farhi, J. Goldstone, S. Gutmann, and M. Sipser, Quantum Computation by Adiabatic Evolution (2000), arXiv:quant-ph/0001106 [quant-ph]

    2000 arXiv

  2. [10]

    Das and B

    A. Das and B. K. Chakrabarti, Colloquium: Quantum annealing and analog quantum computation, Rev. Mod. Phys. 80, 1061 (2008)

    2008

  3. [11]

    Albash and D

    T. Albash and D. A. Lidar, Adiabatic quantum compu- tation, Rev. Mod. Phys. 90, 015002 (2018)

    2018

  4. [12]

    Hauke, H

    P. Hauke, H. G. Katzgraber, W. Lechner, H. Nishi- mori, and W. D. Oliver, Perspectives of quantum anneal- ing: methods and implementations, Rep. Prog. Phys. 83, 054401 (2020)

    2020

  5. [13]

    Cerezo, A

    M. Cerezo, A. Arrasmith, R. Babbush, S. C. Benjamin, S. Endo, K. Fujii, J. R. McClean, K. Mitarai, X. Yuan, L. Cincio, and P. J. Coles, Variational quantum algo- rithms, Nat. Rev. Phys. 3, 625 (2021)

    2021

  6. [14]

    Bharti, A

    K. Bharti, A. Cervera-Lierta, T. H. Kyaw, T. Haug, S. Alperin-Lea, A. Anand, M. Degroote, H. Heimonen, J. S. Kottmann, T. Menke, W.-K. Mok, S. Sim, L.-C. Kwek, and A. Aspuru-Guzik, Noisy intermediate-scale quantum algorithms, Rev. Mod. Phys. 94, 015004 (2022)

    2022

  7. [15]

    Tilly, H

    J. Tilly, H. Chen, S. Cao, D. Picozzi, K. Setia, Y. Li, E. Grant, L. Wossnig, I. Rungger, G. H. Booth, and J. Tennyson, The Variational Quantum Eigensolver: A review of methods and best practices, Phys. Rep. 986, 1 6 (2022)

    2022

  8. [16]

    Farhi, J

    E. Farhi, J. Goldstone, and S. Gutmann, A Quan- tum Approximate Optimization Algorithm (2014), arXiv:1411.4028 [quant-ph]

    2014 arXiv

  9. [17]

    Zhou, S.-T

    L. Zhou, S.-T. Wang, S. Choi, H. Pichler, and M. D. Lukin, Quantum Approximate Optimization Algorithm: Performance, Mechanism, and Implementation on Near- Term Devices, Phys. Rev. X 10, 021067 (2020)

    2020

  10. [18]

    Perez-Garcia, F

    D. Perez-Garcia, F. Verstraete, M. M. Wolf, and J. I. Cirac, Matrix product state representations, Quantum Info. Comput. 7, 401 (2007)

    2007

  11. [19]

    Verstraete, V

    F. Verstraete, V. Murg, and J. I. Cirac, Matrix prod- uct states, projected entangled pair states, and varia- tional renormalization group methods for quantum spin systems, Adv. Phys. 57, 143 (2008)

    2008

  12. [20]

    Schollwöck, The density-matrix renormalization group in the age of matrix product states, Ann

    U. Schollwöck, The density-matrix renormalization group in the age of matrix product states, Ann. Phys. (N. Y.) 326, 96 (2011) , january 2011 Special Issue

    2011

  13. [21]

    Orús, A practical introduction to tensor networks: Matrix product states and projected entangled pair states, Ann

    R. Orús, A practical introduction to tensor networks: Matrix product states and projected entangled pair states, Ann. Phys. (N. Y.) 349, 117 (2014)

    2014

  14. [22]

    M. C. Bañuls, Tensor Network Algorithms: A Route Map, Annu. Rev. Condens. Matter Phys. 14, 173 (2023)

    2023

  15. [23]

    Y. Zhou, E. M. Stoudenmire, and X. Waintal, What Lim- its the Simulation of Quantum Computers?, Phys. Rev. X 10, 041038 (2020)

    2020

  16. [24]

    Ayral, T

    T. Ayral, T. Louvet, Y. Zhou, C. Lambert, E. M. Stoudenmire, and X. Waintal, Density-Matrix Renormal- ization Group Algorithm for Simulating Quantum Cir- cuits with a Finite Fidelity, PRX Quantum 4, 020304 (2023)

    2023

  17. [25]

    R. M. Karp, Reducibility among Combinatorial Prob- lems, in Complexity of Computer Computations. The IBM Research Symposia Series. , edited by R. E. Miller, J. W. Thatcher, and J. D. Bohlinger (Springer US, Boston, MA, 1972) pp. 85–103

    1972

  18. [26]

    M. X. Goemans and D. P. Williamson, Improved approx- imation algorithms for maximum cut and satisfiability problems using semidefinite programming, J. ACM 42, 1115 (1995)

    1995

  19. [27]

    Bowles, A

    J. Bowles, A. Dauphin, P. Huembeli, J. Martinez, and A. Acín, Quadratic Unconstrained Binary Optimization via Quantum-Inspired Annealing, Phys. Rev. Appl. 18, 034016 (2022)

    2022

  20. [28]

    Guaita, L

    T. Guaita, L. Hackl, T. Shi, E. Demler, and J. I. Cirac, Generalization of group-theoretic coherent states for vari- ational calculations, Phys. Rev. Res. 3, 023090 (2021)

    2021

  21. [30]

    Fioroni and V

    L. Fioroni and V. Savona, Entanglement-assisted vari- ational algorithm for discrete optimization problems, Commun. Phys. 8, 438 (2025)

    2025

  22. [31]

    See Supplemental Material at [URL will be inserted by publisher] for additional information on the bench- mark optimization methods, hyperparameter selection, and runtimes

  23. [32]

    Ye, Gset dataset, https://web.stanford.edu/~yyye/ yyye/Gset/ (2003)

    Y. Ye, Gset dataset, https://web.stanford.edu/~yyye/ yyye/Gset/ (2003)

    2003

  24. [33]

    M. C. Bañuls, R. Orús, J. I. Latorre, A. Pérez, and P. Ruiz-Femenía, Simulation of many-qubit quantum computation with matrix product states, Phys. Rev. A 73, 022344 (2006)

    2006

  25. [34]

    J. A. Smolin and G. Smith, Classical signature of quan- tum annealing, Front. Phys. 2, 1 (2014)

    2014

  26. [35]

    Bauer, L

    B. Bauer, L. Wang, I. Pižorn, and M. Troyer, Entangle- ment as a resource in adiabatic quantum optimization (2015), arXiv:1501.06914 [cond-mat.dis-nn]

    2015 arXiv

  27. [36]

    Hatomura and T

    T. Hatomura and T. Mori, Shortcuts to adiabatic classi- cal spin dynamics mimicking quantum annealing, Phys. Rev. E 98, 032136 (2018)

    2018

  28. [37]

    Mugel, C

    S. Mugel, C. Kuchkovsky, E. Sánchez, S. Fernández- Lorenzo, J. Luis-Hita, E. Lizaso, and R. Orús, Dynamic portfolio optimization with real datasets using quantum processors and quantum-inspired tensor networks, Phys. Rev. Res. 4, 013006 (2022)

    2022

  29. [38]

    M. T. Veszeli and G. Vattay, Mean field approximation for solving QUBO problems, PLOS ONE 17, 1 (2022)

    2022

  30. [39]

    G. Lami, P. Torta, G. E. Santoro, and M. Collura, Quan- tum annealing for neural network optimization problems: A new approach via tensor network simulations, SciPost Phys. 14, 117 (2023)

    2023

  31. [40]

    Lopez-Piqueres, J

    J. Lopez-Piqueres, J. Chen, and A. Perdomo-Ortiz, Sym- metric tensor networks for generative modeling and con- strained combinatorial optimization, Mach. Learn.: Sci. Technol. 4, 035009 (2023)

    2023

  32. [41]

    Alcazar, M

    J. Alcazar, M. Ghazi Vakili, C. B. Kalayci, and A. Perdomo-Ortiz, Enhancing combinatorial optimiza- tion with classical and quantum generative models, Nat. Commun. 15, 2761 (2024)

    2024

  33. [42]

    Lopez-Piqueres and J

    J. Lopez-Piqueres and J. Chen, Cons-training tensor net- works: Embedding and optimization over discrete linear constraints, SciPost Phys. 18, 192 (2025)

    2025

  34. [43]

    Nakada, K

    H. Nakada, K. Tanahashi, and S. Tanaka, Quick design of feasible tensor networks for constrained combinatorial optimization, Quantum 9, 1799 (2025)

    2025

  35. [44]

    Rattacaso, D

    D. Rattacaso, D. Jaschke, M. Ballarin, I. Siloi, and S. Montangero, Quantum algorithms for equational rea- soning (2025), arXiv:2508.21122 [quant-ph]

    2025 arXiv

  36. [45]

    García-Sáez and J

    A. García-Sáez and J. I. Latorre, An exact tensor network for the 3SAT problem, Quantum Info. Comput. 12, 283 (2012)

    2012

  37. [46]

    J. D. Biamonte, J. Morton, and J. Turner, Tensor Net- work Contractions for #SAT, J. Stat. Phys. 160, 1389 (2015)

    2015

  38. [47]

    Zhu and H

    Z. Zhu and H. G. Katzgraber, Do tensor renormalization group methods work for frustrated spin systems? (2019), arXiv:1903.07721 [cond-mat.dis-nn]

    2019 arXiv

  39. [48]

    Kourtis, C

    S. Kourtis, C. Chamon, E. R. Mucciolo, and A. E. Ruckenstein, Fast counting with tensor networks, SciPost Phys. 7, 060 (2019)

    2019

  40. [49]

    J.-G. Liu, L. Wang, and P. Zhang, Tropical Tensor Net- work for Ground States of Spin Glasses, Phys. Rev. Lett. 126, 090506 (2021)

    2021

  41. [50]

    M. M. Rams, M. Mohseni, D. Eppens, K. Jałowiecki, and B. Gardas, Approximate optimization, sampling, and spin-glass droplet discovery with tensor networks, Phys. Rev. E 104, 025308 (2021)

    2021

  42. [51]

    T. Hao, X. Huang, C. Jia, and C. Peng, A Quantum- Inspired Tensor Network Algorithm for Constrained Combinatorial Optimization Problems, Front. Phys. 10, 1 (2022)

    2022

  43. [52]

    J.-G. Liu, X. Gao, M. Cain, M. D. Lukin, and S.-T. Wang, Computing Solution Space Properties of Combi- natorial Optimization Problems Via Generic Tensor Net- works, SIAM J. Sci. Comput. 45, A1239 (2023)

    2023

  44. [53]

    Pancotti and J

    N. Pancotti and J. Gray, One-step replica symmetry 7 breaking in the language of tensor networks (2023), arXiv:2306.15004 [quant-ph]

    2023 arXiv

  45. [54]

    Yasuda, S

    S. Yasuda, S. Sotobayashi, and Y. Minato, HOBOTAN: Efficient Higher Order Binary Optimization Solver with Tensor Networks and PyTorch (2024), arXiv:2407.19987 [cs.MS]

    2024 arXiv

  46. [55]

    Tesoro, I

    M. Tesoro, I. Siloi, D. Jaschke, G. Magnifico, and S. Montangero, Quantum inspired factorization up to 100-bit RSA number in polynomial time (2024), arXiv:2410.16355 [cs.CR]

    2024

  47. [56]

    A. A. Gangat and J. Gray, Hyperoptimized approxi- mate contraction of tensor networks for rugged-energy- landscape spin glasses on periodic square and cubic lat- tices, Phys. Rev. E 110, 065306 (2024)

    2024

  48. [57]

    Patra, S

    S. Patra, S. Singh, and R. Orús, Projected entangled pair states with flexible geometry, Phys. Rev. Res. 7, L012002 (2025)

    2025

  49. [58]

    A. M. Ali, Explicit Solution Equation for Every Com- binatorial Problem via Tensor Networks: MeLoCoToN (2025), arXiv:2502.05981 [cs.ET]

    2025 arXiv

  50. [59]

    A. M. Dziubyna, T. Śmierzchalski, B. Gardas, M. M. Rams, and M. Mohseni, Limitations of tensor-network approaches for optimization and sampling: A compari- son to quantum and classical Ising machines, Phys. Rev. Appl. 23, 054049 (2025)

    2025

  51. [60]

    Fishman, S

    M. Fishman, S. R. White, and E. M. Stoudenmire, The ITensor Software Library for Tensor Network Calcula- tions, SciPost Phys. Codebases , 4 (2022)

    2022

  52. [61]

    Fishman, S

    M. Fishman, S. R. White, and E. M. Stoudenmire, Code- base release 0.3 for ITensor, SciPost Phys. Codebases , 4 (2022)

    2022

  53. [62]

    Fröwis, V

    F. Fröwis, V. Nebendahl, and W. Dür, Tensor operators: Constructions and applications for long-range interaction systems, Phys. Rev. A 81, 062337 (2010)

    2010

  54. [63]

    Barahona, M

    F. Barahona, M. Grötschel, M. Jünger, and G. Reinelt, An Application of Combinatorial Optimization to Statis- tical Physics and Circuit Layout Design, Oper. Res. 36, 493 (1988)

    1988

  55. [64]

    Lucas, Ising formulations of many NP problems, Front

    A. Lucas, Ising formulations of many NP problems, Front. Phys. 2, 1 (2014)

    2014

  56. [65]

    Sreedhar, P

    R. Sreedhar, P. Vikstål, M. Svensson, A. Ask, G. Johans- son, and L. García-Álvarez, The Quantum Approximate Optimization Algorithm performance with low entangle- ment and high circuit depth (2022), arXiv:2207.03404 [quant-ph]

    2022 arXiv

  57. [66]

    Chertkov, G

    A. Chertkov, G. Ryzhakov, G. Novikov, and I. Oseledets, Optimization of Functions Given in the Tensor Train For- mat (2022), arXiv:2209.14808 [math.NA]

    2022 arXiv

  58. [67]

    best”, “avg

    W. H. Press, S. A. Teukolsky, W. T. Vetterling, and B. P. Flannery, Numerical Recipes 3rd Edition (Cam- bridge University Press, 2007). End Matter Additional QiILS results for w3R graphs —Here, we evaluate QiILS’s performance across w3R graphs of sizes n = 50, 100, 150, and 20...

    2007

  59. [68]

    H. R. Lourenço, O. C. Martin, and T. Stützle, Iterated Local Search: Framework and Applications, in Handbook of Meta- heuristics, edited by M. Gendreau and J.-Y. Potvin (Springer International Publishing, Cham, 2019) pp. 129–168

    2019

  60. [69]

    Bowles, A

    J. Bowles, A. Dauphin, P. Huembeli, J. Martinez, and A. Acín, Quadratic Unconstrained Binary Optimization via Quantum- Inspired Annealing, Phys. Rev. Appl. 18, 034016 (2022)

    2022

  61. [70]

    Fioroni and V

    L. Fioroni and V. Savona, Entanglement-assisted variational algorithm for discrete optimization problems, Commun. Phys. 8, 438 (2025)

    2025

  62. [71]

    Guaita, L

    T. Guaita, L. Hackl, T. Shi, E. Demler, and J. I. Cirac, Generalization of group-theoretic coherent states for variational calculations, Phys. Rev. Res. 3, 023090 (2021)

    2021

  63. [72]

    P. M. Schindler, T. Guaita, T. Shi, E. Demler, and J. I. Cirac, Variational Ansatz for the Ground State of the Quantum Sherrington-Kirkpatrick Model, Phys. Rev. Lett. 129, 220401 (2022)

    2022

  64. [73]

    Ye, Gset dataset, https://web.stanford.edu/~yyye/yyye/Gset/ (2003)

    Y. Ye, Gset dataset, https://web.stanford.edu/~yyye/yyye/Gset/ (2003). S5 p sweeps iterations λ Fig. 1 (QiILS) 0.2 200 – – Fig. 1 (ILS) 0.03 200 – 1.0 Fig. 2 (a) 0.5 80 1,2,4,8 – Fig. 2 (b) 0.5 80 – 0.5 Fig. 2 (b) (inset) – 80 1 0.5 Fig. 2 (c) 0.1,0.3,0.5,0.7 80 – 0.5 Fig. 2 (...

    2003

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.