Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Quantum Solvers: Predictive Aeroacoustic & Aerodynamic modeling

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

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

Pith's one-line read This paper claims that computational fluid dynamics problems can be exponentially compressed by representing fields and operators as tensor networks, and demonstrates the compression on a point-source wave equation and on steady flow…

desk verdict A competent challenge report with one solid externally benchmarked result and one unsupported exponential-resource claim; archival value only, and the follow-up papers are the citable versions. read the letter →

arxiv 2507.21683 v1 pith:Q5LW2OMT submitted 2025-07-29 quant-ph

Nis-Luca van Hülst , Theofanis Panagos , Greta Sophie Reese , Shahram Panahiyan , Tomohiro Hashizume This is my paper
classification quant-ph
keywords matrixproductstatesoperatorstensornetworksvariationalquantumalgorithmcomputationalfluiddynamicsaeroacousticscurvilinearcoordinatesNavier-Stokesequations
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 claims that the fields and operators of computational fluid dynamics can be carried in tensor-network form whose width stays roughly constant as the computational grid grows, converting what would otherwise be an exponential cost in the number of grid points into a polynomial one. To make this concrete, it demonstrates two extensions of an earlier MPS-MPO framework. First, matrix product operators that advance a pressure field are translated into quantum gates inside a variational quantum algorithm (TPVQA), and the resulting emulated quantum computation reproduces the analytic solution for a 2D wave equation driven by a point source. Second, the incompressible Navier-Stokes equations are written in curvilinear coordinates and solved with matrix product states, predicting the steady drag on a cylinder with a relative error below 0.1% against an industrial finite-volume reference while compressing the solution by more than a factor of 10. The report argues that if the needed bond dimensions stay bounded, the same machinery enables larger-scale CFD simulations on both classical tensor-network hardware and future quantum computers.

What carries the argument

The load-bearing objects are matrix product states and matrix product operators: a field on $2^n$ grid points is amplitude-encoded as a chain of tensors with bond dimension $\chi$, and differential or time-stepping operators are represented as operator chains with bond dimension $\zeta$. Three pieces of machinery carry the argument. The curvilinear-coordinate transformation computes grid metrics by acting with Cartesian derivative MPOs on MPS-encoded grid coordinates, so geometry enters the operators themselves. The projection (fractional-step) time integrator for the incompressible Navier-Stokes equations splits each step into a convective update, a Poisson solve for pressure, and a divergence-free projection, each performed as tensor contractions that the paper scales as $O(\chi^3)$. TPVQA reshapes each MPO tensor into a unitary by padding the bond index to dimension $2^z$ and optimizing under an isometric constraint, then evaluates the variational cost function with one ancilla qubit via phase kickback, so the measurement cost is $O(1+\lceil\log_2 \zeta\rceil)$ qubits and does not grow with the discretization size.

What would settle it

Run the MPS-based curvilinear solver on the same cylinder at a Reynolds number high enough to produce unsteady vortex shedding, fix a target error, and record the bond dimension required as the grid size is increased; if the bond dimension grows with grid size, the compression ratio stops improving and the claimed exponential resource reduction fails.

Watch

Extended reading notes

Core claim

The core discovery is that two different bottlenecks in computational fluid dynamics both admit low-rank tensor representations, and that the same operator format bridges classical tensor-network solvers and quantum circuits. For the aeroacoustic case, the time-stepping operator of the discretized wave equation is written as a matrix product operator whose bond dimension depends only on the finite-difference stencil order, not on the grid size; TPVQA reshapes each MPO tensor into a unitary on a padded Hilbert space and evaluates the variational cost function through a single-ancilla phase-kickback measurement. On a noise-free emulated quantum computer the resulting pressure field at $t = 0.1$ s matches the analytic point-source solution, and the fidelity inside the domain improves as the stencil order increases from 2 to 6. For the aerodynamic case, the curvilinear coordinate transform enters through metrics built from MPS representations of the grid coordinates, and the projection (fractional-step) method is executed with MPS fields and MPO operators at total complexity $O(\chi^3)$, where $\chi$ is the maximum bond dimension. The steady pressure coefficient on a cylinder with $n = 18$ qubits per spatial dimension and $\chi = 35$ yields a horizontal force of roughly $0.7097$ N, within $0.1\%$ of the industrial reference, while using only $8.4\%$ of the variables a direct simulation needs; the measured compression ratio improves with grid size in a way the paper reads as exponential.

Load-bearing premise

The claimed exponential savings rest on the assumption that the tensor-network bond dimension of the flow fields and of the curvilinear operators stays small and nearly independent of the grid size; the report verifies this only for a steady laminar cylinder flow at effective Reynolds number 20 with bond dimension 35, and for a point-source wave field, not for unsteady, turbulent, or higher-Reynolds flows or complex geometries.

Editorial extensions

If this is right

  • If the bond dimensions stay bounded, the same MPS-MPO toolbox applies to any finite-difference PDE, nonlinear, higher-dimensional, or on body-fitted grids, because the operators admit low-rank forms that do not depend on the grid size.
  • TPVQA removes the need to hand-design quantum circuits for CFD: given the MPO of a time-stepping operator, the circuit follows automatically, and each expectation value in the cost function costs only $O(1+\lceil\log_2 \zeta\rceil)$ measured qubits.
  • Higher-order finite-difference stencils cost only $O(\kappa)$ extra MPO bond dimension and $O(\log_2 \kappa)$ extra auxiliary qubits, so derivative accuracy can be raised substantially at modest resource overhead.
  • The curvilinear solver can be switched to a compressible solver and to other geometries without changing the $\chi$-scaling, and the report demonstrates body-fitted grid generation around an aerofoil as a step toward that goal.
  • Accurate steady-state quantities such as the drag force survive aggressive truncation of the transient dynamics, so the method can target engineering figures of merit directly instead of resolving every intermediate detail.

Reading between the lines

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

  • A test the report does not perform is unsteady flow: recording the bond dimension needed to hold a fixed error as the grid grows for vortex shedding at Reynolds numbers above the steady laminar regime would probe whether the exponential compression survives turbulent dynamics.
  • The TPVQA demonstration is limited to a linear wave equation; applying the same MPO-to-unitary translation to one nonlinear projection step would show whether the translation error and the variational optimization remain controlled when operators contain products of fields.
  • The resource counts assume states with area-law entanglement, the kind MPS represents well; flows with volume-law entanglement would lose compression, so the method's practical regime is likely confined to flows dominated by coherent structures, which is what the benchmark cases show.
  • Because the compression ratio improves with grid size in the cylinder case, the same solver could act as a pre-screener for design optimization: coarse tensor-network solves could rank candidate geometries cheaply before expensive high-resolution CFD is run.
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 is a Phase II competition submission reproduced verbatim. It proposes (i) TPVQA, which translates MPO time-stepping operators into quantum gates and evolves a 2D linearized-Euler/wave equation with a point source on an emulated noise-free quantum computer, matching the analytic far-field amplitude; and (ii) an MPS/MPO curvilinear-coordinate Chorin solver for steady incompressible flow around a cylinder, benchmarked against an OpenFOAM reference, reporting a drag-force error below 0.1% and a compression factor exceeding 10 at n=18, χ=35. The abstract and conclusions claim an exponential reduction in computational resources for CFD.

Significance. The claimed exponential compression, if it held across flow regimes and geometries, would be a substantial advance for quantum-inspired CFD. The paper is creditable for benchmarking against an external analytic solution (Case 1) and an industrial-grade CFD reference (Case 2), and for explicitly stating its idealizations (exact expectation values, global optimizer, ansatz containing the solution). However, the evidence for the central exponential-resource claim is confined to one steady laminar cylinder and a point-source wave field; no growth law for χ as a function of grid size, Reynolds number, or geometry is provided. The contribution is therefore promising but not yet established at the level claimed.

major comments (3)
  1. [Section III.B, Fig. 4c, Conclusion] The exponential-resource claim in the Summary and Conclusion rests on the premise that the MPS bond dimension χ remains small and roughly independent of grid size and flow complexity, giving O(2nχ²) parameters instead of 2^(2n). The presented evidence is a single steady laminar cylinder at effective Re≈20 with fixed χ=35, the compression curve in Fig. 4c for that configuration, and the near-constant bond dimension of the curvilinear Laplacian on the same cylinder grid (Fig. 4d). The compression curve is a power-law fit over three values of n (14, 16, 18) and is based on one integrated observable, the horizontal force; a low drag error does not by itself bound the field error. No evidence is offered for unsteady separation, higher Reynolds numbers, or the NACA0012 geometry, for which only a grid is shown. If χ grows with n or with flow complexity, the O(χ³) per-step complexity yields no asymptotic speedup and the claimed exponential reduction fails. The paper should either demonstrate a growth law χ_ε(n, Re) over a range of regimes and geometries or explicitly restrict the exponential claim to the demonstrated cases.
  2. [App. E1, Fig. 3b] The fidelity analysis in Fig. 3b is not a physical benchmark: the 'numerically exact' state used for comparison is generated by applying the same approximate operators Qκ1 and Q2 in exact matrix-vector form, as stated in App. E1. Consequently, the reported fidelities and mean absolute errors isolate the MPO-to-unitary translation error and the finite-difference stencil error, but they do not validate the wave-equation discretization or the sponge layer against the true solution. The analytic far-field comparison in Fig. 1b is the actual external validation. The main text and figure caption should label the fidelity curves as translation-error diagnostics, so that readers do not mistake them for physical accuracy measures.
  3. [Section I, App. C] The manuscript is not self-contained for reproducibility. Section I declares that Phase 1 [8] is 'an integral part of the manuscript' and is assumed to have been read, but [8] is only available from the authors upon reasonable request. Load-bearing statements, including the MPO representation of the sponge operator with bond dimension 4 (App. C), the low-rank MPO representations of the finite-difference operators, and the construction of the initial states, are either cited to unpublished preprints [9] or delegated to Phase 1. If this arXiv record is to serve as a standalone technical report, the essential derivations should be included or the dependency on inaccessible material clearly flagged in the abstract.
minor comments (5)
  1. [Section III.A, App. E2] The Summary states that TPVQA 'only requir[es] to measure O(1+⌈log2 ζ⌉) qubits,' while the protocol in App. E2 evaluates each cost term with a single ancilla readout and z=4 auxiliary qubits; these statements should be reconciled and the measurement cost stated precisely.
  2. [Section II] 'point spurce' should be 'point source.'
  3. [Fig. 3 caption] The labels 'Purity' and 'Mean absolute error' appear merged into one line for panels (c) and (d); please separate them for clarity.
  4. [App. F2] The scaling of the Poisson solve is given as O(χ^4) for a DMRG-type approach and then O(χ³) via the method of Oseledets and Dolgov; the main text cites only O(χ³). Please state explicitly which algorithm is implemented in the reported simulations.
  5. [App. C, Ref. [9]] Reference [9] is cited as 'In preparation' for the sponge operator MPO bond dimension; please provide a public reference or include the derivation in an appendix.

Circularity Check

1 steps flagged · score 2.0 of 10

No significant circularity: external benchmarks carry the physical claims; one internal 'numerically exact' fidelity check is self-referential but non-load-bearing, and self-citations are minor.

  1. self definitional [Appendix E 1 (numerically exact evaluation); Fig. 3b caption]
    "For the evaluation of the numerically exact evolution, |p(x, y, t)⟩exact, we used the same operators Qκ1 and Q2 in the numerically exact form ... Fidelities between the evolved state and numerically exact result excluding the boundary sponge layer after imposing the normalization Finside."

    The 'numerically exact' reference used for the fidelity benchmark in Fig. 3b is constructed by applying the same approximate MPO-derived operators Qκ1 and Q2 that the TPVQA evolution is approximating. The fidelity therefore reduces to a self-consistency check of the variational optimization against its own operator inputs; it does not benchmark the physical accuracy of the discretization or of the MPO-to-unitary translation. This label is overclaimed, but the step is not load-bearing because the physical validation in Fig. 1b compares against the analytic Hankel-function solution.

full rationale

The paper's headline claims rest on two external benchmarks: the analytic 2D wave-equation solution (App. E3) for TPVQA case 1 and the OpenFOAM reference for the cylinder drag (Section III.B). Neither benchmark is produced by fitting a free parameter; χ=35 is a convergence choice from Fig. 4a, not a fit to the drag datum, and the 0.1% drag agreement is a genuine external check. The exponential-compression statement is a definitional property of the MPS/MPO parameterization (O(nχ²) variables vs O(2^n) grid points) conditional on χ staying bounded; the paper supplies support for bounded χ only for its specific cylinder grid and point-source field (Figs. 4c,d), so the general claim is an extrapolation and a stated limitation rather than a circular derivation. Self-citations to Phase 1 and the companion preprints [1,2] are frequent, but the low-rank MPO facts are also grounded in independent literature (e.g., Refs. [37,40,43,44]) and in the paper's own demonstrations, so the citation chain is not the sole load-bearing support. The only self-referential element is the 'numerically exact' fidelity reference computed with the same operators, which is a minor internal-consistency issue, not the basis of the paper's main physical claims. Overall, the derivation is not circular at the level of its external validations; score 2 reflects the minor self-referential benchmark and the weight of self-citation, not any constructional equivalence between prediction and input.

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

The ledger shows the main external costs are not hidden constants but structural assumptions: bounded bond dimension, ideal VQA behavior, and unpublished reference [9]. The listed fitted numbers are standard numerical hyperparameters rather than parameters tuned to match the benchmark outputs.

free parameters (5)
  • Maximum MPS bond dimension chi = 35 for benchmark; swept 20, 30, 40, 50, 60
    Truncation rank that sets compression versus accuracy. The headline 0.1% force error uses chi=35, and Fig. 4a shows accuracy saturating only for chi around 50, so the chosen value is an important numerical choice.
  • Finite difference stencil order kappa = 2, 4, 6 for case 1; second order for case 2
    Discretization order controls the translation error in TPVQA; the paper reports higher kappa improves fidelity but increases MPO bond dimension linearly.
  • Time step Delta t = 1e-4 s for case 1; not stated for case 2
    Explicit time-stepping step in Eq. 3 and in Chorin's method. Stability and accuracy depend on it, and no convergence study is given for case 2.
  • Sponge layer hyperparameters = not stated explicitly
    gamma_max, c_b, and layer extents in App. A are hand-chosen for absorbing boundaries; the low-rank MPO bond dimension of 4 is attributed to unpublished reference [9].
  • MPO-to-unitary translation tolerance epsilon = 3e-7 to 5e-8
    Stopping tolerance for the Riemannian optimization in App. D; it bounds the translation error of the quantum operators but is not independently verified.
assumptions (5)
  • domain assumption Flow fields and operators of interest admit MPS/MPO representations with bond dimension chi that stays small and roughly independent of grid size n.
    This is the core compression premise. Demonstrated for the steady laminar cylinder and point-source cases, but not proven for general CFD.
  • domain assumption For TPVQA, the variational ansatz contains the solution, the measured expectation values are exact, and the classical optimizer reaches the global optimum.
    Stated explicitly in Section III.A Results. Without these idealizations the quantum resource counts and accuracy claims are not validated.
  • ad hoc to paper The MPO-to-unitary translation converges to the reported tolerances with a non-vanishing success probability.
    The translation uses Riemannian optimization from Termanova et al. and unpublished reference [9]; Fig. 3a shows non-vanishing success probability for the specific operators used here, but no general guarantee is given.
  • ad hoc to paper The sponge operator gamma has an accurate MPO representation with bond dimension 4.
    Claimed in App. C with reference to unpublished [9]; this underlies the low-rank treatment of absorbing boundaries.
  • standard math The analytical far-field amplitude formula (App. E3) is valid for the simulated point-source regime.
    Used as the external benchmark for the TPVQA evolution; the formula is cited to Ostashev et al. and Rienstra.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Quantum Solvers: Predictive Aeroacoustic & Aerodynamic modeling." pith.science (2026). https://pith.science/paper/Q5LW2OMT

@misc{pith2026250721683,
  author       = {Pith},
  title        = {Pith review of: Quantum Solvers: Predictive Aeroacoustic & Aerodynamic modeling},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Q5LW2OMT}},
  note         = {Machine review of arXiv:2507.21683}
}
read the original abstract

This technical report presents our winning contribution to the 2024 Airbus and BMW Group Quantum Computing Challenge under the category 'Quantum Solvers'. This submission addresses efficient simulation in industrial CFD using (i) quantum-inspired algorithms and (ii) hybrid quantum-classical algorithms. We reproduce the submitted materials exactly as handed in, providing an archival record, with the sole addition of a note citing the publication that resulted from this challenge.

Figures

Figures reproduced from arXiv: 2507.21683 by the authors.

Figure 1
Figure 1. Case I: a. Translation of matrix-vector op￾eration to a quantum circuit. In Phase 1, a field |u⟩ and an operator O in matrix-vector form (top) is compressed into MPSs and MPOs (yellow and green, bottom right). In phase 2 (bottom left), we translate of MPO to a quantum circuit UO for use in the realm of TPVQA that solves for variational problem parameterized by ⃗θ. Here, the unitary U( ⃗θ) encodes the time evolution … view at source ↗
Figure 2
Figure 2. a. MPO tensors to unitaries. j th unitary Uj of dimension 2 × 2 z is recovered by treating the combined left bond and an incoming physical leg of the j th tensor of Q, Q[j], as the incoming dimension of Uj (red), and the combined right bond and outgoing physical leg of Q[j] as the outgoing dimension of Uj (blue). After projecting out the auxiliary dimension in Q (black box), Q is recovered (light blue). b. VQA with … view at source ↗
Figure 3
Figure 3. a. Evolution of success probabilities (Psucc). Shown are Psucc for evaluating Re{⟨p(t + ∆t)| Q1 |p(t)⟩} (blue) and Re{⟨p(t + ∆t)| Q2 |p(t − ∆t)⟩} (red). b. Fideli￾ties. Fidelities between the evolved state and numerically exact result excluding the boundary sponge layer after im￾posing the normalization Finside. c. Mean absolute error. Mean absolute error of an evolved state to the numerically exact result inside th… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Simulation of a flow around a cylinder: a. Relative errors of the simulations. ϵp (top) and ε|u| (bottom) between MPS-based and DNS simulations as a func￾tion of the time steps for different bond dimensions of 20, 30, 40, 50, and 60 (light to dark), respectively, for a…
Figure 5
Figure 5. Figure 5: Depiction of curvilinear coordinates. We define our curvilinear coordinates parametrized by ξ and η in the physical domain (left) in such a way that the η = 0 boundary of the orthogonal computational grid (right) corresponds to the boundary of an immersed object (red).…

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Quantum-Inspired Simulation of 2D Turbulent Rayleigh-B\'enard Convection

    physics.flu-dyn 2026-04 unverdicted novelty 7.0 of 10

    Matrix product state simulations of 2D Rayleigh-Bénard convection recover Nusselt number statistics with 1.8% error and a 9-fold reduction in degrees of freedom at Ra=10^10 using bond dimensions comparable to lower Ra cases.

Reference graph

Works this paper leans on

54 extracted references · 41 canonical work pages · cited by 1 Pith paper

  1. [8]

    N. L. van H¨ ulst, G. S. Reese, T. Panagos, O. Negrete, S. Panahiyan, and T. Hashizume, Quantum solvers: Predictive aeroacoustic & aerodynamic modeling (phase 1), (2024), available from the authors upon reasonable request

    work page 2024

  2. [9]

    Siegl, G

    P. Siegl, G. S. Reese, T. Hashizume, N. van H¨ ulst, and D. Jaksch, Full quantum tensor scheme for differential equations, In preparation

    work page

  3. [1]

    Siegl, G

    P. Siegl, G. S. Reese, T. Hashizume, N.-L. van H¨ ulst, and D. Jaksch, Tensor-programmable quantum circuits for solving differential equations (2025), arXiv:2502.04425 [quant-ph]

    arXiv 2025

  4. [2]

    van H¨ ulst, P

    N.-L. van H¨ ulst, P. Siegl, P. Over, S. Bengoechea, T. Hashizume, M. G. Cecile, T. Rung, and D. Jaksch, Quantum- inspired tensor-network fractional-step method for incompressible flow in curvilinear coordinates (2025), arXiv:2507.05222 [physics.flu-dyn]

    arXiv 2025

  5. [3]

    Jaksch, J

    D. Jaksch, J. I. Cirac, P. Zoller, S. L. Rolston, R. Cˆ ot´ e, and M. D. Lukin, Fast quantum gates for neutral atoms, Phys. Rev. Lett. 85, 2208 (2000)

    2000

  6. [4]

    Bluvstein, H

    D. Bluvstein, H. Levine, G. Semeghini, T. T. Wang, S. Ebadi, M. Kalinowski, A. Keesling, N. Maskara, H. Pichler, M. Greiner, V. Vuleti´ c, and M. D. Lukin, A quantum processor based on coherent transport of entangled atom arrays, Nature 604, 451 (2022)

    work page 2022

  7. [5]

    Bluvstein, S

    D. Bluvstein, S. J. Evered, A. A. Geim, S. H. Li, H. Zhou, T. Manovitz, S. Ebadi, M. Cain, M. Kalinowski, D. Hangleiter, J. P. Bonilla Ataides, N. Maskara, I. Cong, X. Gao, P. Sales Rodriguez, T. Karolyshyn, G. Semeghini, M. J. Gullans, M. Greiner, V. Vuleti´ c, and M. D. Lukin, Logical quantum processor based on reconfigurable atom arrays, Nature626, 58 (2024)

    2024

  8. [6]

    S. J. Evered, D. Bluvstein, M. Kalinowski, S. Ebadi, T. Manovitz, H. Zhou, S. H. Li, A. A. Geim, T. T. Wang, N. Maskara, H. Levine, G. Semeghini, M. Greiner, V. Vuleti´ c, and M. D. Lukin, High-fidelity parallel entangling gates on a neutral-atom quantum computer, Nature 622, 268 (2023)

    work page 2023

Show all 54 references
  1. [7]

    G. Q. AI and Collaborators, Quantum error correction below the surface code threshold (2024), arXiv:2408.13687

    2024 arXiv

  2. [10]

    Schollwoeck, The density-matrix renormalization group in the age of matrix product states, Annals of Physics 326, 96 (2011), 1008.3477 [cond-mat]

    U. Schollwoeck, The density-matrix renormalization group in the age of matrix product states, Annals of Physics 326, 96 (2011), 1008.3477 [cond-mat]

    2011 arXiv

  3. [11]

    Termanova, A

    A. Termanova, A. Melnikov, E. Mamenchikov, N. Belokonev, S. Dolgov, A. Berezutskii, R. Ellerbrock, C. Mansell, and M. Perelshtein, Tensor Quantum Programming (2024)

    2024

  4. [12]

    Peruzzo, J

    A. Peruzzo, J. McClean, P. Shadbolt, M.-H. Yung, X.-Q. Zhou, P. J. Love, A. Aspuru-Guzik, and J. L. O’Brien, A variational eigenvalue solver on a photonic quantum processor, Nature Communications 5, 4213 (2014)

    2014

  5. [13]

    J. R. McClean, J. Romero, R. Babbush, and A. Aspuru-Guzik, The theory of variational hybrid quantum-classical algo- rithms, New Journal of Physics 18, 023023 (2016)

    2016

  6. [14]

    McArdle, T

    S. McArdle, T. Jones, S. Endo, Y. Li, S. C. Benjamin, and X. Yuan, Variational ansatz-based quantum simulation of imaginary time evolution, npj Quantum Information 5, 1 (2019)

    2019

  7. [15]

    Lubasch, J

    M. Lubasch, J. Joo, P. Moinier, M. Kiffner, and D. Jaksch, Variational quantum algorithms for nonlinear problems, Physical Review A 101, 010301 (2020)

    2020

  8. [16]

    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 algorithms, Nature Reviews Physics 3, 625 (2021)

    2021

  9. [17]

    C. L. Ladson, Effects of independent variation of Mach and Reynolds numbers on the low-speed aerodynamic characteristics of the NACA 0012 airfoil section , Vol. 4074 (National Aeronautics and Space Administration, Scientific and Technical . . . , 1988)

    1988

  10. [18]

    J. Russell, Crossing the Quantum Threshold: The Path to 10,000 Qubits — hpcwire.com, https://www.hpcwire.com/ 2024/04/15/crossing-the-quantum-threshold-the-path-to-10000-qubits/ , [Accessed 30-10-2024]

    2024

  11. [19]

    Israeli and S

    M. Israeli and S. A. Orszag, Approximation of radiation boundary conditions, Journal of Computational Physics 41, 115 (1981)

    1981

  12. [20]

    B. A. Mani, On the reflectivity of sponge zones in compressible flow simulations (2010)

    2010

  13. [21]

    Peri´ c and M

    R. Peri´ c and M. Abdel-Maksoud, Analytical prediction of reflection coefficients for wave absorbing layers in flow simulations of regular free-surface waves, Ocean Engineering 147, 132 (2018)

    2018

  14. [22]

    Carmigniani and D

    R. Carmigniani and D. Violeau, Optimal sponge layer for water waves numerical models, Ocean Engineering 163, 169 (2018)

    2018

  15. [23]

    Here, for the sake of brevity, we use p(t) instead of p(x, y, t)

  16. [24]

    Aharonov, V

    D. Aharonov, V. Jones, and Z. Landau, A Polynomial Quantum Algorithm for Approximating the Jones Polynomial, Algorithmica 55, 395 (2009)

    2009

  17. [25]

    Nielsen and I

    M. Nielsen and I. Chuang, Quantum Computation and Quantum Information: 10th Anniversary Edition (Cambridge University Press, 2010)

    2010

  18. [26]

    de Frutos, V

    J. de Frutos, V. John, and J. Novo, Projection methods for incompressible flow problems with weno finite difference schemes, Journal of Computational Physics 309, 368 (2016). 10

    2016

  19. [27]

    Y. Zang, R. L. Street, and J. R. Koseff, A non-staggered grid, fractional step method for time-dependent incompressible navier-stokes equations in curvilinear coordinates, Journal of Computational Physics 114, 18 (1994)

    1994

  20. [28]

    J. H. Ferziger and M. Peri´ c,Computational Methods for Fluid Dynamics , 3rd ed. (Springer, Berlin, 2002)

    2002

  21. [29]

    Anderson, Computational Fluid Dynamics: The Basics with Applications , McGraw-Hill International Editions: Mechan- ical Engineering (McGraw-Hill, 1995)

    J. Anderson, Computational Fluid Dynamics: The Basics with Applications , McGraw-Hill International Editions: Mechan- ical Engineering (McGraw-Hill, 1995)

    1995

  22. [30]

    A. J. Chorin, Numerical solution of the navier-stokes equations, Mathematics of Computation 22, 745 (1968)

    1968

  23. [31]

    A. A. Michailidis, C. Fenton, and M. Kiffner, Tensor train multiplication (2024), arXiv:2410.19747 [physics.comp-ph]

    2024 arXiv

  24. [32]

    I. V. Oseledets and S. V. Dolgov, Solution of linear systems and matrix inversion in the tt-format, SIAM Journal on Scientific Computing 34, A2718 (2012), https://doi.org/10.1137/110833142

    2012 doi

  25. [33]

    Paeckel, T

    S. Paeckel, T. K¨ ohler, A. Swoboda, S. R. Manmana, U. Schollw¨ ock, and C. Hubig, Time-evolution methods for matrix- product states, Annals of Physics 411, 167998 (2019), arxiv:1901.05824 [cond-mat, physics:quant-ph]

    2019 arXiv

  26. [34]

    W. J. Gordon and L. C. Thiel, Transfinite mappings and their application to grid generation, Applied Mathematics and Computation 10-11, 171 (1982)

    1982

  27. [35]

    H. G. Weller, G. Tabor, H. Jasak, and C. Fureby, A tensorial approach to computational continuum me- chanics using object-oriented techniques, Computer in Physics 12, 620 (1998), https://pubs.aip.org/aip/cip/article- pdf/12/6/620/7865493/620 1 online.pdf

    1998

  28. [36]

    Clancy, Aerodynamics (Longman Scientific & Technical, 1986)

    L. Clancy, Aerodynamics (Longman Scientific & Technical, 1986)

    1986

  29. [37]

    For the MPS, the NVPS is calculated by summing the sizes of its individual tensors, of order O(2nχ2), while for the DNS it is given by NVPS DNS = 2n

    and compare its value for the MPS with that of DNS. For the MPS, the NVPS is calculated by summing the sizes of its individual tensors, of order O(2nχ2), while for the DNS it is given by NVPS DNS = 2n. The results are presented in Fig. 4c. Notably, for n = 18, which repre- sen...

  30. [38]

    Kiffner and D

    M. Kiffner and D. Jaksch, Tensor network reduced order models for wall-bounded flows (2023), arxiv:2303.03010 [physics, physics:quant-ph]

    2023 arXiv

  31. [39]

    R. L. Sorenson, A computer program to generate two-dimensional grids about airfoils and other shapes by the use of poisson’s equation (1980)

    1980

  32. [40]

    J. F. Thompson, Elliptic grid generation, Applied Mathematics and Computation 10-11, 79 (1982)

    1982

  33. [41]

    Gourianov, M

    N. Gourianov, M. Lubasch, S. Dolgov, Q. Y. van den Berg, H. Babaee, P. Givi, M. Kiffner, and D. Jaksch, A quantum- inspired approach to exploit turbulence structures, Nature Computational Science 2, 30 (2022)

    2022

  34. [42]

    Oseledets, Constructive representation of functions in low-rank tensor formats, Constructive Approximation 37 (2010)

    I. Oseledets, Constructive representation of functions in low-rank tensor formats, Constructive Approximation 37 (2010)

    2010

  35. [43]

    MITCHELL and D

    A. MITCHELL and D. Griffiths, The Finite Difference Method in Partial Differential Equations (1980)

    1980

  36. [44]

    Kazeev and B

    V. Kazeev and B. Khoromskij, Low-Rank Explicit QTT Representation of the Laplace Operator and Its Inverse, SIAM Journal on Matrix Analysis and Applications 33, 10.1137/100820479 (2012)

    2012 doi

  37. [45]

    J. J. Garc ´ ıa-Ripoll, Quantum-inspired algorithms for multivariate analysis: from interpolation to partial differential equa- tions, Quantum 5, 431 (2021)

    2021

  38. [46]

    Nibbi and C

    M. Nibbi and C. B. Mendl, Block encoding of matrix product operators (2024), arXiv:2312.08861 [quant-ph]

    2024 arXiv

  39. [47]

    J. Li, F. Li, and S. Todorovic, Efficient riemannian optimization on the stiefel manifold via the cayley transform, CoRR abs/2002.01113 (2020), 2002.01113

    2020 arXiv

  40. [48]

    V. E. Ostashev, D. K. Wilson, L. Liu, D. F. Aldridge, N. P. Symons, and D. Marlin, Equations for finite-difference, time- domain simulation of sound propagation in moving inhomogeneous media and numerical implementation, The Journal of the Acoustical Society of America 117, 503 (2005)

    2005

  41. [49]

    S. W. Rienstra, A class of cylindrically symmetric exact solutions of the wave equation, Journal of Sound and Vibration 545, 117430 (2023). 11 Appendix A: Sponge operators for absorbing boundary conditions We define the sponge operator γ(x, y) = γ(x) + γ(y) for implementing ab...

    2023

  42. [50]

    Cost function The evolution of an unnormalized pressure field by one-time step, |p(x, y, t+ ∆t)⟩ to |p(x, y, t+ ∆t)⟩ = λp(t + ∆t)U (⃗θopt(t + ∆t)) |0⟩⊗n, is achieved by optimizing a parameterized circuit U (⃗θ) with a set of classical parameters θ0 and ⃗θ = {θ1, θ2, . . .}. Wi...

  43. [51]

    We evaluate the expectation value of a Hermitianized unitary operator Q† + Q by measuring the σz expectation value of the ancilla qubit

    Hadamard test For the efficient evaluation of the expectation values in a cost function, we utilize the phase kickback algorithm [15, 24] by introducing one more ancillary qubit. We evaluate the expectation value of a Hermitianized unitary operator Q† + Q by measuring the σz e...

  44. [52]

    2, where boundary is at infinity is analytically solvable, and is known to be analytically expressible in terms of Hankel functions and describes outgoing waves [47]

    Analytical result for 2D wave equation with a point source The wave equation Eq. 2, where boundary is at infinity is analytically solvable, and is known to be analytically expressible in terms of Hankel functions and describes outgoing waves [47]. Using polar coordinates, in t...

  45. [53]

    Here, ρ refers to the density, ν to the kinematic viscosity

    Chorin’s projection The incompressible Navier-Stokes equations are given by [26] ∇ ·u = 0, (F1) ∂tu + (u · ∇)u = − ∇p ρ + ν∆u, (F2) where u = (u, v) is the velocity field and p the pressure. Here, ρ refers to the density, ν to the kinematic viscosity. To solve this initial val...

  46. [54]

    This step involves contractions of MPO-MPS in a pointwise manner

    Scaling Firstly, we have an offline step, the transformation step, which needs to be performed once per system and grid, to create the metrics and curvilinear differential operators. This step involves contractions of MPO-MPS in a pointwise manner. Hence, it scales as O(χ2 met...

Pith tools

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