Pith. sign in

REVIEW 3 major objections 5 minor 35 references

Solving Differential Equations Using Continuous-Variable Quantum Annealing

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read A linear differential equation can be discretized with finite differences and turned into a quadratic cost over continuous variables; the paper claims that continuous-variable quantum annealing on bosonic modes finds the ground state…

desk verdict A clean but thin proof of principle: the ODE-to-CV-QA mapping is correct and the substitution is exact for this Hamiltonian class, but the proof and numerics are under-reported. read the letter →

arxiv 2608.04601 v1 pith:PYRP2Z3E submitted 2026-08-05 quant-ph

classification quant-ph
keywords continuous-variablequantumannealinglineardifferentialequationsbosonicmodesfinite-differencemethodquadraticoptimizationcavityadiabaticevolution
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 paper tries to establish that linear differential equations can be solved by continuous-variable quantum annealing, in which each unknown grid value is a bosonic mode rather than a register of qubits. The authors discretize the equation with finite differences, square the residual to get a quadratic cost, and map that cost to a Hamiltonian whose ground state's quadrature expectations give the solution. Numerical simulation of two second-order linear ordinary differential equations with fixed endpoints reproduces the analytical solutions under the tested conditions. The significance would be a route to differential-equation solvers that avoid the precision-driven qubit overhead of binary encoding.

What carries the argument

The central object is the operator substitution rule of Eq. (4), which converts the quadratic cost $L(\theta)=\theta^T Q\theta + r^T\theta + c$ into a bosonic Hamiltonian built from number operators $a_i^\dagger a_i$, beamsplitter-style couplings $(a_i^\dagger a_j + a_i a_j^\dagger)/2$, and linear drives $(a_i+a_i^\dagger)/2$. The finite-difference matrix $C$ from the differential equation enters through $Q=C^T C$ and $r=-2C^T b$, so the classical residual is exactly the quadratic cost. The driver Hamiltonian $H_D=d\sum_i a_i^\dagger a_i$ has the vacuum as its ground state, and adiabatic evolution from vacuum to $H_P$ is intended to leave the system in $H_P$'s ground state, whose quadrature expectations are read out as $x(\tau_i)$.

What would settle it

Compare the ground-state quadrature readouts against the classical least-squares solution for a discretized equation where the classical residual has a unique minimum and the Hamiltonian is known to have a squeezed ground state; a detectable deviation beyond the finite-difference error would show the mapping does not preserve the optimization problem.

Watch

Extended reading notes

Core claim

The central claim is that the squared residual $L(x)=(Cx-b)^T(Cx-b)$ of a finite-difference discretization of a linear differential equation is a quadratic form, and applying the operator substitution $\theta_i^2 \to a_i^\dagger a_i$, $\theta_i\theta_j \to (a_i^\dagger a_j + a_i a_j^\dagger)/2$, $\theta_i \to (a_i+a_i^\dagger)/2$ yields a bosonic problem Hamiltonian whose ground state, read out through $\langle (a_i+a_i^\dagger)/2\rangle$, solves the discretized equation. The paper demonstrates this for $d^2x/d\tau^2=2$ and $d^2x/d\tau^2+5\,dx/d\tau=-1$ with $x(0)=0$, $x(1)=1$, using $M=7$ grid intervals, annealing time $T=10000$, and a numerical simulation library; in both cases the annealed values are close to the analytical solutions.

Load-bearing premise

The method assumes that the quantum ground state of the bosonic Hamiltonian, after the variable-to-operator substitution, has average readings that match the minimum of the original classical equation residual; this is inherited from earlier work and is not proved for differential equations here.

Editorial extensions

If this is right

  • A linear ordinary differential equation of any finite order can be turned into a quadratic cost by finite differences, so the construction is not limited to the two example equations.
  • Because each variable is encoded in one bosonic mode, the number of physical modes equals the number of unknown grid points, rather than the number of grid points times the bits per variable.
  • Increasing the annealing time $T$ drives the final expectation value of $H_P$ toward zero, meaning longer annealing improves satisfaction of the discretized equations.
  • Under conditions where annealing reaches the ground state, the results are relatively insensitive to the driver strength $d$.
  • The two tested second-order equations reproduce their analytical solutions at $M=7$ and $T=10000$, indicating that a moderate grid can suffice for these cases.

Reading between the lines

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

  • A direct extension the paper does not spell out is that the same squared-residual construction applies to systems of coupled linear differential equations, since the finite-difference matrix $C$ simply grows to cover all variables and the cost remains quadratic.
  • The numerical trend that $\langle H_P\rangle$ stops approaching zero as $M$ grows suggests that the bosonic encoding trades qubit count for sensitivity to adiabaticity; the practical overhead may reappear as longer annealing times rather than disappearing.
  • Because the substitution rule omits the $a^2$ and $a^{\dagger 2}$ terms that an exact quantization of $\theta_i^2$ would contain, a squeezed or non-coherent ground state could in principle spoil the readout; checking the second moments of the final state would test reliability beyond the two examples.
  • For nonlinear differential equations the residual is no longer quadratic, so the method as stated does not extend; whether a polynomial residual could be mapped by higher-order operator products is an open question the paper leaves implicit.
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 / 5 minor

Summary. The manuscript extends continuous-variable quantum annealing (CV-QA) to linear ordinary differential equations. The authors discretize the time interval, approximate derivatives with central finite differences to obtain a linear system Cx=b, square the residual to produce the quadratic cost L(x)=(Cx-b)^T(Cx-b), and map this cost to a bosonic problem Hamiltonian using the substitution rules of Eq. (4) with quadrature readout. QuTiP simulations for two second-order boundary-value problems show that for M=5,7,9 and sufficiently long annealing times, the readout quadratures appear to agree with the analytic solutions and the final problem-Hamiltonian expectation decreases with T. The paper concludes that this establishes a potential route to solving differential equations without the qubit overhead of binary encodings.

Significance. The central derivation is transparent and the matrices C and b for the two examples are correct; the residual formulation in Eq. (12) is a clean least-squares approach. If the ground-state readout claim is proven, the result would be a useful proof-of-principle extension of CV-QA from linear regression (Ref. [18]) to differential equations, potentially avoiding the precision-dependent qubit overhead of QUBO-based approaches. The paper is honest about the restricted conditions examined. The current significance is limited by two gaps: no proof that the ground state of the constructed Hamiltonian has quadrature expectations minimizing L(x), and numerical evidence that is only qualitative (plots plus a residual whose definition is ambiguous). Both are fixable in revision.

major comments (3)
  1. [Sec. II A, Eq. (4); Sec. III, Eq. (23)] The paper does not prove that the ground state of the problem Hamiltonian has quadrature expectations minimizing L(x); it simply inherits the substitution rules from Ref. [18]. For the Hamiltonians actually constructed here this is true and should be shown: H_P is a quadratic, number-conserving bosonic Hamiltonian with only linear drive terms and no pairing terms, so it can be completed to H_P = (a + 1/2 Q^{-1} r)^\dagger Q (a + 1/2 Q^{-1} r) - (1/4) r^T Q^{-1} r + c, whose ground state is a multi-mode coherent state with \langle x_i \rangle = - (1/2)(Q^{-1} r)_i = (C^T C)^{-1} C^T b. The authors should add this argument, discuss the case where Q is singular (pseudo-inverse, possible ground-state degeneracy), and clarify that the general substitution Eq. (4) is not an exact quantization for arbitrary cost functions. Without this proof, the central claim rests entirely on two numerical examples.
  2. [Sec. IV B, Eq. (23); Figs. 2, 3, 5, 6] The quantity plotted as the expectation value of \hat H_P is inconsistent with the definition in Eq. (23), which includes the additive constant c = b^T b from Eq. (14). For Eq. (16) with M=7, c is approximately 2229, so \langle \hat H_P \rangle cannot approach zero if Eq. (23) is used literally. If the simulation omitted the constant term, as is standard for optimization, the paper should state this explicitly and define the residual operator actually plotted; otherwise the convergence criterion is ill-posed. This matters because Figs. 2 and 5 are the main quantitative evidence for convergence.
  3. [Sec. IV A and Figs. 1-6] The numerical results are not sufficiently specified or quantified. The paper states that simulations were implemented in QuTiP but does not report the Fock-space truncation dimension, the solver or tolerance settings, or a convergence check in the truncation dimension. It also gives no quantitative error between the final \langle x_i \rangle and the analytic solution, e.g., max or RMS deviation over the grid for each M and T; Figs. 1 and 4 show only visual agreement. Please add such error metrics, report the truncation dimension, and state that the closed-system evolution is deterministic, in which case error bars are not required. Without these additions, the conclusion that the solutions converge toward the analytical solutions cannot be independently checked.
minor comments (5)
  1. [Sec. IV B] The phrase 'Section 3' should be 'Section III' for consistent nomenclature.
  2. [Abstract and Sec. I] The acronym CV-QA is used in the abstract and introduction but is never defined; define it at first use.
  3. [Figs. 1 and 4 captions] In the version provided, the captions contain garbled text rendering 'analytic solution'; please check the production files before submission.
  4. [Sec. IV A] A short discussion of why M=7 is sufficient and how the discretization error behaves with M would help separate the finite-difference error from the annealing error.
  5. [Sec. II A, Eq. (5)] Since Q = C^T C is symmetric by construction, the symmetrized coefficient (Q_ij + Q_ji)/2 in Eq. (5) could be simplified to Q_ij, and the paper can note this.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the cost function is derived directly from the differential equation and the numerical benchmarks are against independent analytic solutions.

full rationale

The derivation chain is self-contained at the level that matters for circularity. The linear ODE (Eq. 6) is discretized by finite differences (Eq. 8) to give Cx = b (Eq. 9), and the squared residual L(x) = (Cx-b)^T(Cx-b) (Eq. 12) is formed directly from the residual, with Q = C^T C and r = -2C^T b (Eq. 14). No parameter is fitted to the target solutions; the only free simulation parameters are the driver strength d and annealing time T, which are fixed and then varied to study convergence (Figs. 2-3, 5-6). The operator mapping of Eq. (4) is inherited from Ref. [18], whose authors include two of the present authors, but this is a non-load-bearing self-citation: the substitution is stated explicitly in the paper, is independent of the differential-equation content, and the claim that ground-state quadratures solve the least-squares problem is a property of the resulting quadratic bosonic Hamiltonian, not an input. The numerical results are benchmarked against independent analytic solutions x = τ^2 and x = 6/(5(e^{-5}-1))(e^{-5τ}-1) - τ/5; there is no renaming of a known result and no fitted input renamed as prediction. The most serious weakness is that the paper omits the proof that the ground state of H_P has the required coherent displacement, and it does not report Fock-space truncation or quantitative readout errors; these are correctness/reproducibility concerns, not circularity. Therefore the circularity score is 0.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the CV-QA operator mapping from Ref. [18], on the standard adiabatic theorem, and on finite-difference discretization. No free parameters are fitted to the analytic solutions; M, T, d, and the unstated Fock truncation are hand-chosen simulation settings. No new physical entities are introduced.

free parameters (4)
  • Number of time divisions M = 7 (plus 5 and 9 in tests)
    Chosen by hand; the demonstration and convergence behavior depend on this discretization parameter.
  • Annealing time T = 10000
    Chosen by hand; the final energy approaches zero as T increases, but no target accuracy is used to set T.
  • Driver Hamiltonian strength d = 1.0
    Chosen by hand; the paper shows the result is broadly insensitive to d in the successful regime, but this still sets the annealing schedule.
  • Fock-space truncation dimension = not reported
    QuTiP simulations of bosonic modes require a finite Hilbert-space cutoff; the reported curves depend on this unreported number.
assumptions (5)
  • domain assumption The operator substitution of Eq. (4) preserves the optimization problem: the ground state of the resulting Hamiltonian has quadrature expectations that minimize L(theta).
    Invoked in Section II.A; the exact quantum operator for theta_i^2 would include a_i^2 and a_i^dagger^2 terms, so this is a modeling assumption inherited from Ref. [18], not proven here.
  • standard math The quantum adiabatic theorem allows the vacuum ground state of the driver to evolve into the ground state of the problem Hamiltonian.
    Used in Section II; the required spectral gap and sufficiently slow T are not quantified for the example Hamiltonians.
  • domain assumption Central finite differences give an adequate discretization of the derivatives in Eq. (6).
    Used in Section III; the discretization error is not compared with the analytic solution, and convergence as M grows is not demonstrated at fixed accuracy.
  • domain assumption The discrete system Cx=b has a unique solution for the boundary-value problems considered.
    Required for the squared residual to have a unique minimum; plausible for the two examples, but not verified in the paper.
  • domain assumption The solution value is read out as the expectation value of the quadrature operator x_i = (a_i + a_i^dagger)/2.
    Section IV.A; for coherent states this equals the classical variable, but for arbitrary states the expectation value need not minimize L(theta).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Solving Differential Equations Using Continuous-Variable Quantum Annealing." pith.science (2026). https://pith.science/paper/PYRP2Z3E

@misc{pith2026260804601,
  author       = {Pith},
  title        = {Pith review of: Solving Differential Equations Using Continuous-Variable Quantum Annealing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PYRP2Z3E}},
  note         = {Machine review of arXiv:2608.04601}
}
read the original abstract

Most existing quantum annealing approaches are formulated for qubit-based architectures. Consequently, applying them to continuous-variable optimization problems requires discretizing the variables, which can incur substantial qubit overhead. Continuous-variable quantum annealing based on bosonic systems has recently been proposed as an alternative framework, in which each optimization variable is directly encoded in a bosonic mode, such as a cavity mode. In this work, we develop a continuous-variable quantum annealing formulation for solving linear differential equations. By recasting the determination of the solution as a continuous-variable optimization problem, the differential equation can be mapped onto an objective function compatible with bosonic quantum annealing. Numerical simulations of second-order linear differential equations demonstrate that, under the conditions considered, the proposed formulation reproduces the corresponding analytical solutions. These results establish a potential route toward solving differential equations without the discretization overhead inherent in qubit-based implementations.

Figures

Figures reproduced from arXiv: 2608.04601 by the authors.

Figure 1
Figure 1. compares the solution obtained by the pro￾posed method with the analytical solution x(τ) = τ 2 . The values of x obtained by QA are close to the analyt￾ical solution. Although T = 10000 is used here, under the present conditions this annealing time was sufficient to obtain a solution close to the analytical one. The number of divisions is set to M = 7, and this moder￾ate number of divisions was sufficient to reprodu… view at source ↗
Figure 2
Figure 2. FIG. 2. Relationship between the expectation value of [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Relationship between the expectation value of [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (2 more)
Figure 5
Figure 5. Figure 5: FIG. 5. Relationship between the expectation value of [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6. Relationship between the expectation value of [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 29 canonical work pages

  1. [18]

    Karimi and P

    S. Karimi and P. Ronagh, Practical integer-to-binary mapping for quantum annealers, Quantum Information Processing 18, 94 (2019)

  2. [1]

    Kadowaki and H

    T. Kadowaki and H. Nishimori, Quantum annealing in the transverse ising model, Physical Review E 58, 5355–5363 (1998)

  3. [2]

    The elements of b depend on the constant term B(τ ) in Eq

    and on the derivative order n of the variable x. The elements of b depend on the constant term B(τ ) in Eq. ( 7): b = (B(τ0), B(τ1), · · · , B(τM ))T (11) The squared residual of Eq. ( 9) is then L(x) = ( Cx − b)T (Cx − b) (12) and expanding Eq. ( 12) gives L(x) = xT CT Cx − 2bT Cx + bT b (13) Defining Q = CT C r = −2CT b, c = bT b (14) we obtain L(x) = xT...

  4. [3]

    Farhi, J

    E. Farhi, J. Goldstone, S. Gutmann, and M. Sipser, Quantum computation by adiabatic evolution, arXiv preprint quant-ph/0001106 (2000)

  5. [4]

    Farhi, J

    E. Farhi, J. Goldstone, S. Gutmann, J. Lapan, A. Lund- gren, and D. Preda, A quantum adiabatic evolution al- gorithm applied to random instances of an np-complete problem, Science 292, 472 (2001)

  6. [5]

    2, increasing T makes the expectation value of ˆHP approach zero

    As in Fig. 2, increasing T makes the expectation value of ˆHP approach zero. Thus, even when both first- and second-derivative terms are included, the same qualitative convergence behavior is observed. Con- versely, when T is on the order of several hundred, the 6 expectation value of ˆHP does not sufficiently converge to zero. Figure 6 shows that, in the re...

  7. [6]

    A. B. Finnila, M. A. Gomez, C. Sebenik, C. Stenson, and J. D. Doll, Quantum annealing: A new method for mini- mizing multidimensional functions, Chemical physics let- ters 219, 343 (1994)

  8. [7]

    In this work, we use central finite differences as a concrete approximation

    becomes a quadratic func- tion of x(τi). In this work, we use central finite differences as a concrete approximation. Examples are dx dτ ≃ x(τi+1) − x(τi−1) 2∆τ (8) d2x dτ 2 ≃ x(τi+1) − 2x(τi) + x(τi−1) ∆τ 2 d3x dτ 3 ≃ x(τi+2) − 2x(τi+1) + 2x(τi−1) − x(τi−2) 2∆τ 3 . . . Applying the above approximations to the derivative terms in Eq. ( 7), we obtain Cx − b ...

Show all 35 references
  1. [8]

    Kato, On the adiabatic theorem of quantum me- chanics, Journal of the Physical Society of Japan 5, 435 (1950)

    T. Kato, On the adiabatic theorem of quantum me- chanics, Journal of the Physical Society of Japan 5, 435 (1950)

  2. [9]

    ( 13) to Eq

    are given by x = (x(τ1), x(τ2), · · · , x(τM−1))T (20) Cij =            − 2 (∆τ )2 , i = j, 1 (∆τ )2 , |i − j| = 1, 0, otherwise, (21) and b = ( 2 − x(τ0) (∆τ )2 , 2, · · · , 2, 2 − x(τM ) (∆τ )2 ) T (22) By applying the transformation from Eq. ( 13) to Eq. (15), th...

  3. [10]

    Apolloni, C

    B. Apolloni, C. Carvalho, and D. De Falco, Quantum stochastic optimization, Stochastic Processes and their Applications 33, 233 (1989)

  4. [11]

    Lucas, Ising formulations of many NP problems, Fron- tiers in Physics 2, 5 (2014)

    A. Lucas, Ising formulations of many NP problems, Fron- tiers in Physics 2, 5 (2014)

  5. [12]

    Hauke, H

    P. Hauke, H. G. Katzgraber, W. Lechner, H. Nishimori, and W. D. Oliver, Perspectives of quantum annealing: methods and implementations, Reports on Progress in Physics 83, 054401 (2020)

  6. [13]

    Jansen, M.-B

    S. Jansen, M.-B. Ruskai, and R. Seiler, Bounds for the adiabatic approximation with applications to quan- tum computation, Journal of Mathematical Physics 48 (2007)

  7. [14]

    Ambainis and O

    A. Ambainis and O. Regev, An elementary proof of the quantum adiabatic theorem, arXiv preprint quant- ph/0411152 (2006)

  8. [15]

    B. Dema, J. Arai, and K. Horikawa, Support vector ma- chine for multiclass classification using quantum anneal- ers, in Proc. DEIM Forum (2020)

  9. [16]

    Date and T

    P. Date and T. Potok, Adiabatic quantum linear regres- sion, Scientific reports 11, 21905 (2021)

  10. [17]

    Chancellor, Domain wall encoding of discrete variab les for quantum annealing and qaoa, Quantum Science and Technology 4, 045004 (2019)

    N. Chancellor, Domain wall encoding of discrete variab les for quantum annealing and qaoa, Quantum Science and Technology 4, 045004 (2019)

  11. [19]

    Tamura, T

    K. Tamura, T. Shirai, H. Katsura, S. Tanaka, and N. To- gawa, Performance comparison of typical binary-integer encodings in an ising machine, IEEE Access 9, 81032 (2021). 7

  12. [20]

    Krakoff, S

    B. Krakoff, S. M. Mniszewski, and C. F. Negre, Con- trolled precision qubo-based algorithm to compute eigen- vectors of symmetric matrices, Plos one 17, e0267954 (2022)

  13. [21]

    S. Arai, H. Oshiyama, and H. Nishimori, Effectiveness of quantum annealing for continuous-variable optimization, Phys. Rev. A 108, 042403 (2023)

  14. [22]

    Koura, T

    A. Koura, T. Imoto, K. Ura, and Y. Matsuzaki, Linear regression using quantum annealing with continuous vari- ables, Japanese Journal of Applied Physics 64, 03SP31 (2025)

  15. [23]

    Bravo-Prieto, R

    C. Bravo-Prieto, R. LaRose, M. Cerezo, Y. Subasi, L. Cincio, and P. J. Coles, Variational quantum linear solver, Quantum 7, 1188 (2023)

  16. [24]

    Y. Sato, H. C. Watanabe, R. Raymond, R. Kondo, K. Wada, K. Endo, M. Sugawara, and N. Yamamoto, Variational quantum algorithm for generalized eigen- value problems and its application to the finite-element method, Physical Review A 108, 022429 (2023)

  17. [25]

    Liu, Y.-S

    H.-L. Liu, Y.-S. Wu, L.-C. Wan, S.-J. Pan, S.-J. Qin, F. Gao, and Q.-Y. Wen, Variational quantum algorithm for the poisson equation, Phys. Rev. A 104, 022418 (2021)

  18. [26]

    Y. Sato, R. Kondo, S. Koide, H. Takamatsu, and N. Imoto, Variational quantum algorithm based on the minimum potential energy for solving the poisson equa- tion, Physical Review A 104, 052409 (2021)

  19. [27]

    Y. Liu, Z. Chen, C. Shu, S. C. Chew, B. C. Khoo, X. Zhao, and Y. Cui, Application of a variational hybrid quantum-classical algorithm to heat conduction equation and analysis of time complexity, Physics of Fluids 34 (2022)

  20. [28]

    Demirdjian, D

    R. Demirdjian, D. Gunlycke, C. A. Reynolds, J. D. Doyle, and S. Tafur, Variational quantum solutions to the advection–diffusion equation for applications in fluid dy- namics, Quantum Information Processing 21, 322 (2022)

  21. [29]

    Kudo, Annealing-based approach to solving partial differential equations, in Proc

    K. Kudo, Annealing-based approach to solving partial differential equations, in Proc. 2026 Int. Conf. Quantum Communications, Networking, and Computing (QCNC) (IEEE, 2026) pp. 912–917

  22. [30]

    Srivastava and V

    S. Srivastava and V. Sundararaghavan, Box algorithm for the solution of differential equations on a quantum annealer, Phys. Rev. A 99, 052355 (2019)

  23. [31]

    G. G. Pollachini, J. P. L. C. Salazar, C. B. D. G´ oes, T. O. Maciel, and E. I. Duzzioni, Hybrid classical-quantum ap- proach to solve the heat equation using quantum anneal- ers, Phys. Rev. A 104, 032426 (2021)

  24. [32]

    J. C. Criado and M. Spannowsky, Qade: solving differen- tial equations on quantum annealers, Quantum Science and Technology 8, 015021 (2022)

  25. [33]

    Johansson, P

    J. Johansson, P. Nation, and F. Nori, Qutip 2: A python framework for the dynamics of open quan- tum systems, Computer Physics Communications 184, 1234–1240 (2013)

  26. [34]

    Lambert, E

    N. Lambert, E. Gigu‘ere, P. Menczel, B. Li, P. Hopf, G. Su’arez, M. Gali, J. Lishman, R. Gadhvi, R. Agarwal, A. Galicia, N. Shammah, P. Nation, J. R. Johansson, S. Ahmed, S. Cross, A. Pitchford, and F. Nori, Qutip 5: The quantum toolbox in Python, Physics Reports 1153, 1 (2026)

  27. [35]

    Seki and H

    Y. Seki and H. Nishimori, Quantum annealing with an- tiferromagnetic fluctuations, Phys. Rev. E 85, 051112 (2012)

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.