Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

High order schemes for solving partial differential equations on a quantum computer

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read High-order spatial discretization reduces qubit count but not the number of Trotter steps needed for a fixed accuracy.

desk verdict Useful Pauli-decomposition toolbox, but the headline Trotter-step claim rests on an unproven commutativity step. read the letter →

arxiv 2412.19232 v1 pith:4AI34GBG submitted 2024-12-26 quant-ph cs.ET

Boris Arseniev , Dmitry Guskov , Richik Sengupta , Igor Zacharov This is my paper · ORCID
classification quant-phcs.ET MSC 81P6865M0635L05
keywords quantumPDEalgorithmsHamiltoniansimulationTrotterizationhigh-orderfinitedifferencesPaulidecompositiond-bandmatriceswaveequationgatecomplexity
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

This paper asks whether higher-order spatial discretization can make Hamiltonian-based quantum algorithms for partial differential equations cheaper in gate count. The authors build a method for decomposing d-band matrices into Pauli strings grouped into commuting sets, then simulate the one-dimensional wave equation with central finite-difference schemes of orders up to kappa = 10. They find that raising the discretization order reduces the number of qubits needed to reach a target discretization error, just as in classical numerics. They also find, and state as Proposition 3, that the total numerical error is approximately the sum of an independent Trotter error and a discretization error, given by epsilon_ns(t) <= epsilon_tr(t,r) + t c_max epsilon_ds(k). The practical consequence is that the Trotterization gate bottleneck persists even when high-order schemes are used.

What carries the argument

The load-bearing object is the d-band matrix B_k that represents the central finite-difference approximation of d/dx, together with the symmetrized Hamiltonian H_k(c) = (1/h) [[0, B_k(c)], [B_k^T(c), 0]] that turns the wave equation into a Schrodinger evolution. A d-band matrix has non-zero entries on the main diagonal and on d diagonals above and below it. The paper decomposes such a matrix into Pauli strings using the Walsh-function parameterization W(x,z) and groups the strings into internally commuting subsets by the condition x dot z = 0 or 1 mod 2. This decomposition lets a circuit be built from parameterized R_z rotations after Clifford diagonalization. The error analysis then separates the Trotter error, governed by product-formula bounds, from the discretization error epsilon_ds(k) = O($h^{{2k-1/2}}$).

What would settle it

For a variable wave speed such as c(x) = 2 + sin(x), compute the exact evolution error ||$e^{{-itH_{d/dx}}$} - $e^{{-itH_k}}$|| and compare it with the claimed t c_max epsilon_ds(k) bound; if the measured difference exceeds that bound by more than the linearization remainder, the commutativity step in the proof of Proposition 3 is invalid. A simpler observable is to run the same fixed-accuracy Trotter-step search with non-constant c(x) and check whether the required r actually drops as k increases.

Watch

Extended reading notes

Core claim

The paper makes two connected claims. First, for any d-band matrix of size 2^n by 2^n, the Pauli strings that can appear in its decomposition are exactly those indexed by x_{k,j} = MBIN_{n-s}(2j-1) * MBIN_s(2^s - k), with s = ceil(log2 k), and these strings can be grouped into internally commuting subsets by the parity of x dot z mod 2. This gives a closed-form count s(d,n) = $2^{{BINL(d)}}$ + (n - BINL(d))d of sets, enabling circuit construction without explicit Pauli matrix multiplication. Second, numerical experiments on the wave equation with Dirichlet boundary conditions show that higher-order schemes reduce the qubits needed for a given discretization error but do not reduce the Trotter steps needed to keep the total error fixed. Proposition 3 formalizes this as the additive bound epsilon_ns(t) <= epsilon_tr(t,r) + t c_max epsilon_ds(k), with the two error sources acting independently.

Load-bearing premise

The proof of the additive error bound assumes that the continuous wave operator and its discretized matrix version commute, which is automatic for constant wave speed but not generally for a variable one.

Editorial extensions

If this is right

  • For a fixed target accuracy in the one-dimensional wave equation, the number of Trotter steps r is set by the Trotter error rather than by the spatial discretization error, so choosing a higher-order stencil changes r only mildly through the Hamiltonian norm.
  • The Pauli decomposition gives a gate count g = O(r (2n^2 + 2n) s(k,n)) = O(r n k 2^n) for a 2^n by 2^n problem with sparse B_k, avoiding explicit Pauli matrix multiplication in circuit construction.
  • Within a fixed error budget, the Trotter error only needs to be kept at or below that budget; beyond that, the spatial order can be increased without adding Trotter steps, which is the operating point the authors recommend.
  • For high-order central schemes with Dirichlet boundaries, the solution should be extended anti-symmetrically and the wave speed symmetrically into the Hamiltonian; otherwise the boundary treatment degrades the scheme's accuracy.
  • Because the Trotter and discretization errors act additively, improving spatial accuracy alone cannot improve the overall error once the Trotter plateau is reached, so gate reduction must come from the time-integration side.

Reading between the lines

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

  • A consequence the paper leaves implicit is that the resource-allocation rule suggested by Proposition 3, namely set r by the Trotter error and spend the remaining error budget on spatial order, plausibly applies to Hamiltonian-based PDE solvers beyond the one-dimensional wave equation, although only the wave equation is tested.
  • The commuting-subset decomposition of d-band matrices could be reused for other sparse Hamiltonian simulation tasks, such as discretized diffusion or elasticity operators, where the same Trotter bottleneck would likely reappear.
  • If the additive error bound is examined under a non-constant wave speed c(x), the commutativity step in the proof of Proposition 3 is not automatically valid; a numerical check on variable-speed problems would determine whether the no-Trotter-reduction conclusion extends beyond the constant-speed case tested here.
  • Should the additive bound survive variable speeds, the practical takeaway is that future algorithmic improvements should target the time integrator, such as higher-order Trotter formulas or randomized product formulas, rather than the spatial stencil.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes a method for decomposing d-band diagonal matrices into Pauli strings grouped into commuting sets, extending the authors' earlier tridiagonal decomposition [15]. Using this decomposition, the authors simulate the one-dimensional wave equation on a quantum circuit model with central finite-difference spatial discretizations of order 2k. Their main numerical claim is that higher-order spatial discretization reduces the number of qubits needed for a given accuracy but does not reduce the number of Trotter steps required to preserve solution accuracy. Proposition 3 formalizes this as an approximately additive error bound, ϵns(t) ≤ ϵtr(t,r) + t cmax ϵds(k). The paper also reports gate-complexity estimates showing that the Trotter error dominates the asymptotic gate count.

Significance. If the central claim holds, it is practically important: it says that replacing low-order finite differences by high-order ones in Hamiltonian-based quantum PDE solvers buys qubit reduction but not a relaxation of the Trotter-step bottleneck. The proposed decomposition of d-band matrices into internally commuting Pauli subsets is a concrete algorithmic contribution and appears to be a natural extension of the authors' prior work. A notable strength is that the paper provides pseudocode and a GitHub link for the numerical experiments, and the numerical data in Figures 2-4 are clear and reproducible in spirit. However, the formal error decomposition in Proposition 3 rests on a commutativity assertion that is false for the finite-difference operators used, so at present the broad version of the headline claim is supported only by constant-coefficient numerics.

major comments (3)
  1. [Section III, Proposition 1 and Table III] Formula (10) as printed, x_{k,j}=MBIN_{n-s}(2j-1)*MBIN_s(2^s-k), does not generate the sets shown in Table III or in the code in Appendix C.1. For example, with n=3, k=1, s=0, and j=3, the printed formula gives MBIN_3(5)=101, whereas Table III lists x_{1,3}=111 and the code computes ((1<<3)-1)<<0 = 7. The correct closed form is x_{k,j}=MBIN_{n-s}(2^j-1)*MBIN_s(2^s-k) (equivalently (2^j-1)<<s | (2^s-k)), which matches the listing and the table. Since Proposition 1 is the basis for the decomposition method, Corollary 1, and the circuit construction, this formula error should be corrected in the theorem statement; otherwise a reader following the stated proposition obtains incorrect Pauli supports.
  2. [Appendix A.4, proof of Proposition 3] The proof asserts that 'it can easily be checked that the operators H_{d/dx} and H_k commute' and then factors e^{-it(R+H_k)} as e^{-itR} e^{-itH_k}. This assertion is false for the finite-difference operators actually used. With A=D I_c and A_k=B_k I_c, the top-left block of the commutator [H_{d/dx}, H_k] is B_k c^2 D - D(c^2 B_k). This does not vanish in general: D and B_k do not commute on a finite grid with the boundary modifications of Appendix B, and for nonconstant c(x) the term D(c^2 B_k) contains (c^2)' B_k, which has no counterpart in B_k c^2 D. Consequently the factorization and the bound et||R||-1 do not follow, and the additive bound (31) is not established by the given proof. This is load-bearing because Proposition 3 is the formal basis for the conclusion that higher-order spatial schemes do not reduce the required number of Trotter steps.
  3. [Section IV B and Proposition 3] All numerical experiments in Section IV B use the constant wave speed c=1 (l=5, t=1, c=1). Since the commutativity failure identified above is most severe for variable c(x), and since Proposition 3 is the only formal bridge from the numerical constant-coefficient observations to the general statement in the abstract, the general claim 'higher-order methods do not decrease the number of Trotter steps needed to preserve solution accuracy' currently rests on a numerical observation made only for c=1. I would request either a repaired proof of Proposition 3 or a direct numerical check with nonconstant c(x) (for example c(x)=1+0.5 sin(πx/l)) that reports ϵns, ϵtr, and ϵds separately and verifies whether the additive structure persists.
minor comments (6)
  1. [Section II A] Equation (4) writes ϵds(k)=O(√N h^{2k})≈O(h^{2k-1/2}); for clarity, state the constant factors depending on l explicitly, since h=l/(N-1).
  2. [Section IV A 2] Equation (27) is missing parentheses around the factor that is raised to the power p+1; the intended expression appears to be O((2Γ_k 5^{⌊p/2⌋-1} ||H_k|| t)^{p+1} / r^p). This should be corrected for readability.
  3. [Section IV A 1] The sentence 'In general case (-bk,...,-b1,0,b1,...,bk) the upper left corner of matrix B_k(c) ... is given by' is followed by no visible displayed matrix; the manuscript should include the matrix or clarify the reference to Appendix B.
  4. [Section IV B] The captions of Figures 2 and 3 do not specify the discretization orders κ for each curve; adding a legend or listing κ values in the captions would improve verifiability.
  5. [References] Reference [20] is incomplete: it gives only 'Fornberg, SIAM Rev., 685'; the full author list, volume, article number, and year should be supplied.
  6. [Throughout] The spelling 'trotterization' is used inconsistently with 'Trotterization'; please unify capitalization.

Circularity Check

0 steps flagged · score 1.0 of 10

No material circularity: the central Trotter-step and qubit-count claims are read off numerical experiments against an analytic standing-wave benchmark, not fitted from the claimed conclusion. The main vulnerability is a non-circular proof gap in Proposition 3 (an asserted commutation of H_{d/dx} and H_k), plus standard self-citation to the authors' prior decomposition work.

full rationale

The abstract's two claims are derived from numerical experiments in Section IV B. Figure 2 compares direct matrix-exponential evolution e^{-iH_k t} with the analytic solution u_sw(x,t)=sin(pi x/l)cos(pi c t/l) of eq. (17), using l=5, t=1, c=1. Figures 3 and 4 obtain the required Trotter step count r by binary search against prescribed error targets (Section IV B 1: 'We employed binary search to find the minimal number of steps r needed for a specific accuracy based on the chosen discretization context (n, kappa)'), rather than by assuming Proposition 3. Thus the statement that higher-order discretization does not reduce the number of Trotter steps is a measured numerical result, not a fitted parameter renamed as a prediction. The paper does rely on the authors' prior work [15] for the base tridiagonal Pauli decomposition and for commuting-subset grouping (e.g., 'As shown earlier in Proposition 4 of [15]' and 'This result, derived from our earlier research [15]...'). That is ordinary self-citation to a published, parameter-free mathematical result with stated assumptions that do not include the target claim, so it does not make the derivation circular. No equation reduces to its input by construction. The proof of Proposition 3 in Appendix A.4 contains an unsupported assertion 'it can easily be checked that the operators H_{d/dx} and H_k commute', and the subsequent factorization e^{-it(R+H_k)} = e^{-itR} e^{-itH_k} is used to obtain the additive error bound. This is a correctness risk, especially for variable c(x) with the boundary-modified finite-difference B_k, but it is not a circularity: the bound is not assumed as the claim itself. Score 1 reflects the minor self-citation reliance and the lack of an independent proof of the commutativity step.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The central claim rests on standard finite-difference and Hamiltonian-simulation tools, plus two paper-specific assumptions: commutativity of the exact and discretized Hamiltonians in the error bound, and a boundary-condition modification whose generalization is asserted. No new physical entities are introduced.

free parameters (1)
  • Trotter order p = 2
    The Trotter formula order is fixed to p=2 in all numerical experiments. The authors state that p>=4 yields too few steps for detailed comparison, so the central conclusion on Trotter scaling is only demonstrated for p=2.
assumptions (4)
  • ad hoc to paper The operators H_{d/dx} and H_k commute, so e^{-it(R+H_k)} = e^{-itR} e^{-itH_k} in the proof of Proposition 3.
    Stated in Appendix A.4 without proof; generally false for variable coefficient c(x) because d/dx(c\cdot) does not commute with its discretization B_k I_c.
  • domain assumption The wave equation can be reformulated as a Schrodinger evolution with Hamiltonian of the form [[0, B], [B^T, 0]].
    Taken from Costa et al. [8] and used as the foundation of the quantum algorithm in Section IV A.
  • domain assumption Dirichlet boundary conditions are incorporated via anti-symmetric extension of the solution and symmetric extension of the wave speed, leading to modified matrices of the form -B_k(c) B_k(c)^T with adjusted first column and coefficients alpha=beta=sqrt(2).
    Derived explicitly for k=3 in Appendix B, with the statement 'the generalization is straightforward' for arbitrary k. All numerical results depend on this boundary treatment.
  • standard math Central finite difference coefficients for the first derivative are taken from the Fornberg algorithm [20].
    Used to construct the band matrices B_k in Section II A; coefficients are standard and not derived in the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of High order schemes for solving partial differential equations on a quantum computer." pith.science (2026). https://pith.science/paper/4AI34GBG

@misc{pith2026241219232,
  author       = {Pith},
  title        = {Pith review of: High order schemes for solving partial differential equations on a quantum computer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4AI34GBG}},
  note         = {Machine review of arXiv:2412.19232}
}
abstract

We explore the utilization of higher-order discretization techniques in optimizing the gate count needed for quantum computer based solutions of partial differential equations. To accomplish this, we present an efficient approach for decomposing $d$-band diagonal matrices into Pauli strings that are grouped into mutually commuting sets. Using numerical simulations of the one-dimensional wave equation, we show that higher-order methods can reduce the number of qubits necessary for discretization, similar to the classical case, although they do not decrease the number of Trotter steps needed to preserve solution accuracy. This result has important consequences for the practical application of quantum algorithms based on Hamiltonian evolution.

Figures

Figures reproduced from arXiv: 2412.19232 by the authors.

Figure 1
Figure 1. The figure illustrates the decomposition discussed [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 3
Figure 3. Comparison of ϕ⃗V (t) - solution obtained with pro￾posed quantum algorithm with ⃗u(t) = ⃗usw(t)/∥ ⃗u0∥ with fixed Trotter errors of 10−3 , 10−5 , and 10−7 , represented by dif￾ferent marker shapes. Results for 4 and 5 qubits are given with red and green respectively (the total number of qubits is given by n + 1). The Trotter formula is set to order p = 2; the wave equation parameters l = 5, t = 1, and c = 1 [PITH_F… view at source ↗
Figure 2
Figure 2. Comparison of ϕ⃗V (t) - solution obtained as e −iHktψ⃗(t = 0) with ⃗u(t) = ⃗usw(t)/∥ ⃗u0∥ an analytical stand￾ing wave solution at t = 1, with l = 5 and c = 1 for different accuracy orders κ = 2k and different number of qubits for dis￾cretization n (the total number of qubits is given by n + 1). Moreover, [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: The relationship between the number of Trotter [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Structure-Preserving Quantum Simulation of Wave Equations on a Trapped-Ion Processor

    quant-ph 2026-07 conditional novelty 6.0 of 10

    On Quantinuum H2-2, Fourier-based structure-preserving circuits resolve subdomain kinetic-energy dynamics for structured 1D/2D acoustic and Dirac wave problems up to 4096 encoded degrees of freedom with MAE ~0.006–0.024.

Reference graph

Works this paper leans on

44 extracted references · 34 canonical work pages · cited by 1 Pith paper

  1. [15]

    A. M. Childs, J.-P. Liu, and A. Ostrander, Quantum 5, 574 (2021)

    2021

  2. [1]

    Exclude the tensor product symbol in Pauli strings P , e.g., XY ZY stands for X ⊗ Y ⊗ Z ⊗ Y

    work page

  3. [2]

    For example, {P1, P2}⊗2 is equivalent to {P1, P2} ⊗ {P1, P2}, which is the same as {P1 ⊗ P1, P1 ⊗ P2, P2 ⊗ P1, P2 ⊗ P2}

    The notation {P1, P2}⊗n is the n-fold Cartesian product of the set {P1, P2}, with its elements inter- preted as Pauli strings. For example, {P1, P2}⊗2 is equivalent to {P1, P2} ⊗ {P1, P2}, which is the same as {P1 ⊗ P1, P1 ⊗ P2, P2 ⊗ P1, P2 ⊗ P2}. In a similar manner, we represent the product Pk ⊗ Pj in an abbreviated form as PkPj

    work page

  4. [3]

    , xn) is expressed as x = Pn j=1 xj2n−j

    The integer representation of the bit sequence x = ( x1, . . . , xn) is expressed as x = Pn j=1 xj2n−j. That is the most significant bit (MSB) is on the left. This mapping is represented with func- tion MBIN n(x), defined in Table I. For example, MBIN8(18) = 00010010 for n = 8 and x = 18. This aligns with [18] for the wave function k = |k1, . . . , kn⟩ in...

    work page

  5. [4]

    The function BINL() returns the length (num- ber of bits) of binary representation, for example BINL(13) = 4

    work page

  6. [5]

    red subset

    The notation ∥·∥ represents the l2-norm for vectors and the spectral norm for matrices, defined by the largest singular value. Note that for unitary matri- ces U , it holds that ∥U ∥ = 1. Notation Definition P = {I, X, Y, Z} Set of Pauli matrices ⊕ XOR (addition modulo 2) {P1, P2}⊗n n-th Cartesian product xp where x, p∈ B xp = x ⊕ p = x ⊕ p ⊕ 1 x Negation...

    work page

  7. [6]

    The derivative op- erator, formulated in matrix form (as outlined in Section II A), encounters limitations in accurately approximating derivative values near the boundaries

    Dirichlet boundary conditions We consider the application of Dirichlet boundary con- ditions to the problem defined in (16). The derivative op- erator, formulated in matrix form (as outlined in Section II A), encounters limitations in accurately approximating derivative values near the boundaries. This discrepancy arises because the elements within the ve...

    work page

  8. [7]

    This can be accomplished by expressing the Hamil- tonian as a sum of terms that are easy to implement: Hk = PΓk γ=1 Hk,γ [10]

    Time evolution of the Hamiltonian To execute Hamiltonian evolution on a quantum com- puter, it must be decomposed into fundamental opera- tions. This can be accomplished by expressing the Hamil- tonian as a sum of terms that are easy to implement: Hk = PΓk γ=1 Hk,γ [10]. Following earlier research [15], Hk,γ is composed of mutually commuting Pauli strings...

    work page

Show all 44 references
  1. [8]

    We employed binary search to find the minimal number of steps r needed for a specific accuracy based on the chosen discretization context ( n, κ)

    Results for the quantum algorithm The quantum algorithm for simulating Hamiltonian evolution requires executing r Trotterization steps, with the number of steps affecting the precision of the solu- tion. We employed binary search to find the minimal number of steps r needed fo...

  2. [9]

    J. H. Ferziger, M. Peri´ c, and R. L. Street,Computational methods for fluid dynamics (springer, 2019)

    2019

  3. [10]

    L. Chen, J. Huang, L.-Y. Fu, W. Peng, C. Song, and J. Han, Remote Sensing 15, 604 (2023)

    2023

  4. [11]

    Ann´ e, P

    L. Ann´ e, P. Joly, and Q. H. Tran, Computational Geo- sciences 4, 207 (2000)

    2000

  5. [12]

    Liang, Y.-F

    W.-Q. Liang, Y.-F. Wang, and C.-C. Yang, Journal of Geophysics and Engineering 12, 108 (2015)

    2015

  6. [13]

    Lloyd., Science , 1073–1078 (1996)

    S. Lloyd., Science , 1073–1078 (1996)

    1996

  7. [14]

    S. Jin, N. Liu, and Y. Yu, Phys. Rev. A 108, 032603 (2023)

    2023

  8. [16]

    P. C. S. Costa, S. Jordan, and A. Ostrander, Phys. Rev. A 99, 012323 (2019)

    2019

  9. [17]

    Hagan and N

    M. Hagan and N. Wiebe, Quantum 7, 1181 (2023)

    2023

  10. [18]

    D. W. Berry, G. Ahokas, R. Cleve, and B. C. Sanders, Communications in Mathematical Physics 270, 359 (2007)

    2007

  11. [19]

    D. W. Berry, A. M. Childs, R. Cleve, R. Kothari, and R. D. Somma, Phys. Rev. Lett. 114, 090502 (2015)

    2015

  12. [20]

    M. A. Nielsen and I. L. Chuang, Quantum Computation and Quantum Information (Cambridge University Press, New York, 2011)

    2011

  13. [21]

    Jones, arXiv preprint arXiv:2401.16378 (2024)

    T. Jones, arXiv preprint arXiv:2401.16378 (2024)

    2024 arXiv

  14. [22]

    Vidal Romero and J

    S. Vidal Romero and J. Santos-Su´ arez, Quantum Infor- mation Processing 22, 449 (2023)

    2023

  15. [23]

    Arseniev, D

    B. Arseniev, D. Guskov, R. Sengupta, J. Biamonte, and I. Zacharov, Physical Review A 109, 052629 (2024)

    2024

  16. [24]

    H. F. Trotter, Proceedings of the American Mathematical Society 10, 545 (1959)

    1959

  17. [25]

    A. M. Childs, Y. Su, M. C. Tran, N. Wiebe, and S. Zhu, Physical Review X 11, 011020 (2021)

    2021

  18. [26]

    Welch, D

    J. Welch, D. Greenbaum, S. Mostame, and A. Aspuru- Guzik, New Journal of Physics 16, 033040 (2014)

    2014

  19. [27]

    R. J. LeVeque, Finite Difference Methods for Ordinary and Partial Differential Equations (SIAM, 2007)

    2007

  20. [28]

    , 685 (1998)

    Fornberg, SIAM Rev. , 685 (1998)

    1998

  21. [29]

    G. H. Golub and C. F. Van Loan, Matrix computations (JHU press, 2013)

    2013

  22. [30]

    Kawase and K

    Y. Kawase and K. Fujii, Computer Physics Communica- tions 288, 108720 (2023)

    2023

  23. [31]

    Van Den Berg and K

    E. Van Den Berg and K. Temme, Quantum4, 322 (2020)

    2020

  24. [32]

    Open quantum assembly language,

    A. W. Cross, L. S. Bishop, J. A. Smolin, and J. M. Gambetta, “Open quantum assembly language,” (2017), arXiv:1707.03429 [quant-ph]

    2017 arXiv

  25. [33]

    Difference methods for wave motion,

    H. P. Langtangen, “Difference methods for wave motion,” (2016)

    2016

  26. [34]

    Gustafsson, Math

    B. Gustafsson, Math. Comp. 29, 396 (1975)

    1975

  27. [35]

    Zacharov, R

    I. Zacharov, R. Arslanov, M. Gunin, D. Stefonishin, A. Bykov, S. Pavlov, O. Panarin, A. Maliutin, S. Ryko- vanov, and M. Fedorov, Open Engineering 9 (2019), https://doi.org/10.1515/eng—2019—0059

    2019 doi

  28. [36]

    Github page with numerical experi- ments,

    “Github page with numerical experi- ments,” https://github.com/barseniev/ General-Matrix-Decomposition . 12 Appendix A: Proofs

  29. [37]

    For exam- ple, MBIN8(18) results in the bit string 00010010

    The Order of Binary Strings As explained earlier in Section II, we adopt the stan- dard MSB (Most Significant Bit) format for numbers rep- resented by bit strings (MBIN n(x) function). For exam- ple, MBIN8(18) results in the bit string 00010010. This appendix utilizes the LSB ...

  30. [38]

    , an, 0) and b = (b1, b2,

    Proof of Proposition 1 for decomposition of general tridiagonal matrix Two binary strings a = ( a1, a2, . . . , an, 0) and b = (b1, b2, . . . , bn, 0), corresponding to decimal numbers a and b can be added to obtain a binary string c = a + b = (c1, c2, . . . , cn+1), we append...

  31. [39]

    Proof of Proposition 2 for decomposition of the general tridiagonal matrix Proposition 2 (Number of sets in the decomposition) . The total count of sets Sk,j (including S0) in the decom- position of a d-band matrix B ∈ CN ×N , where N = 2n, as described in Proposition 1 is giv...

  32. [40]

    Proof of Proposition 3 for error impact Proposition 3 (Error for solving wave equation) . The bound for numerical solution error ϵns(t) = ⃗ u(t) − ⃗ϕV (t) , where ⃗ u(t) is the exact solution of the wave equation and ⃗ϕV (t) is the numerical solution ob- tained with 2k discret...

  33. [41]

    17 Appendix C: program listings

    This way we indeed can see that the right hand side of wave equation with incorporated boundary conditions can be written as −Bk(c)Bk(c)T . 17 Appendix C: program listings

  34. [42]

    h> // For CHAR BIT and b u i l t i n c l z /∗∗ ∗ @ b r i e f G e n e r a t e s s u b s e t s based on g i v e n p a r a m e t e r s and f i l l s t h e output a r r a y

    Listing to compute formulae in proposition 1 and 2 for the decomposition of a general tridiagonal matrix #i n c l u d e< l i m i t s . h> // For CHAR BIT and b u i l t i n c l z /∗∗ ∗ @ b r i e f G e n e r a t e s s u b s e t s based on g i v e n p a r a m e t e r s and f i l ...

  35. [43]

    h> 18 #i n c l u d e<s t d i n t

    Listing to compute W alsh function, Equation (6) from pair (xk,j , z) #i n c l u d e<s t d i o . h> 18 #i n c l u d e<s t d i n t . h> // C o n s t a n t s r e p r e s e n t i n g P a u l i m a t r i c e s // and t h e i r c o r r e s p o n d i n g i m a g i n a r y s i g n an...

  36. [44]

    ∗ ∗ Parameters : ∗ − num qubits ( i n t ) : The number o f q u b i t s

    Listing to compute decomposition weights, Equation (8) from pair (xk,j , z) #i n c l u d e<cmath > #i n c l u d e<c s t d d e f> // F u n c t i o n p r o t o t y p e s f o r B ( ) , assuming i t i s d e f i n e d e l s e w h e r e d o u b l e B( l o n g k , l o n g m) ; // C o...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.