Pith. sign in

REVIEW 3 major objections 4 minor 32 references

An unfitted boundary algebraic equation method with Calder\'on preconditioning for 2D Stokes flow in irregular geometry

T0 review · 3 major / 4 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read This paper constructs explicit lattice Green's functions for the 2D staggered MAC Stokes operator and shows a boundary-only algebraic system solves irregular-geometry Stokes flow with second-order accuracy.

desk verdict A genuinely new discretize-then-represent Stokes BAE with explicit MAC lattice Green's functions and strong numerics, but the rank-completion step that guarantees solvability is asserted rather than proved. read the letter →

arxiv 2607.21295 v1 pith:JMB6DHNU submitted 2026-07-23 math.NA cs.NAphysics.comp-phphysics.flu-dyn

Wenjun Ying , Qing Xia This is my paper · ORCID
classification math.NAcs.NAphysics.comp-phphysics.flu-dyn MSC 65N3865N0676D07
keywords StokesequationslatticeGreen'sfunctionboundaryalgebraicequationdiscretepotentialtheoryMACschemeCalderónpreconditioningunfittedmeshMoffatteddies
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 establishes that steady 2D Stokes flow on a staggered MAC grid can be solved as a fully discrete boundary algebraic equation, with no fitted mesh, no singular quadrature, and no artificial far-field boundary conditions. It builds explicit velocity and pressure lattice Green's functions from Laplace and biharmonic lattice Green's functions, then represents homogeneous incompressible fields by sources on thin staggered boundary layers, with Dirichlet data imposed at cut points. Sampled-normal rank updates remove the hydrostatic pressure-jump nullspaces, one per disconnected obstacle, and a componentwise Calderón preconditioner built from the scalar Laplace kernel makes the dense boundary system well conditioned. Numerical tests show second-order velocity and pressure convergence, discrete divergence at solver accuracy, and near mesh-independent conditioning for exterior flows.

What carries the argument

The key machinery is the MAC Stokes lattice Green's function pair (S,P), derived in Theorem 2.1 from the Laplace lattice Green's function G and a gauge-fixed biharmonic representative H via the Duffin–Shelly identity. S is a 2x2 tensor of difference operators applied to H; P is a gradient of G. These translation-invariant kernels invert the MAC Stokes operator and supply both homogeneous layer potentials and volume potentials. Around each kernel sit three supporting devices: cut-point interpolation with piecewise-linear hat functions and optional extrapolation to extra exterior nodes; sampled-normal rank updates that project out the hydrostatic nullspace of the single-layer matrix; and a com

What would settle it

Compute the exact nullspace of the uncompleted boundary matrix A for a non-convex or multiple-obstacle geometry at several grid spacings, and compare its null vectors to the sampled-normal vector nΓ (and per-object normals). If any null vector has substantial component away from nΓ, or if the completed system shows a near-zero singular value that does not shrink under refinement, the rank-completion premise fails.

Watch

Extended reading notes

Core claim

The central discovery is a translation-invariant free-space pair of lattice Green's functions for the coupled steady Stokes operator on the staggered MAC grid: the velocity kernel S and pressure kernel P, constructed in closed form from regularized Laplace and biharmonic lattice Green's functions. These kernels satisfy the discrete momentum equations and the discrete divergence constraint, so convolution with layer sources on thin exterior boundary layers produces homogeneous incompressible fields. Dirichlet data are imposed by local tensor-product hat interpolation at cut points, and the resulting boundary-density equation is completed by a sampled-normal rank update that removes one hydros

Load-bearing premise

The load-bearing premise is that the sampled normal vector at cut points truly represents the discrete hydrostatic pressure-jump mode; if this proxy is wrong for some geometry, the rank-completed boundary system remains singular and the entire solve fails.

Editorial extensions

If this is right

  • Second-order velocity and pressure convergence is demonstrated across smooth, multiply connected, narrow-gap, body-forced, and exterior configurations, so the method is accurate without boundary-fitted meshes.
  • The maximum discrete divergence is bounded between 10^-12 and 10^-7 in all tests, confirming that the unfitted closure preserves the MAC incompressibility constraint to solver accuracy.
  • The Calderón preconditioner reduces the condition number by orders of magnitude and gives nearly mesh-independent conditioning for exterior flows, with GMRES iteration counts in the tens to low hundreds.
  • The recovered Moffatt-eddy scale ratios (2.105, 2.113, 2.095 versus the asymptotic 2.10) show that the method resolves multiscale corner structures.
  • Because the lattice kernels depend only on the background grid, they can be reused when geometry changes, and the boundary operator can be reused when only the body force changes, enabling efficient moving-boundary and time-dependent extensions.

Reading between the lines

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

  • If the sampled-normal rank completion faithfully represents the exact discrete hydrostatic null vector (a claim the paper asserts without a rigorous proof), the method should extend to arbitrarily many disjoint obstacles by adding one rank-one update per object; a careful numerical test comparing the sampled normal to the exact algebraic null vector for a non-convex obstacle would settle this.
  • The explicit Stokes lattice Green's function pair opens a natural route to a discrete double-layer formulation, which the paper lists as future work; such a formulation could reduce the boundary system's size and improve conditioning in narrow gaps.
  • The same discretize-then-represent route should transfer to 3D MAC Stokes and to linear elasticity, since the derivation only needs the Laplace and biharmonic lattice Green's functions and the staggered-grid projection; the main challenges would be larger boundary-layer cardinality and a multidirectional sampled-normal completion.
  • For time-dependent Navier–Stokes, the volume-potential/homogeneous-correction split suggests a semi-implicit scheme where nonlinear terms are treated as known volume sources at each step, reusing the same boundary operator.
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 / 4 minor

Summary. The paper develops an unfitted boundary algebraic equation (BAE) method for 2D steady Stokes flow on a MAC grid. It explicitly constructs free-space velocity and pressure lattice Green's functions for the discrete MAC Stokes operator from Laplace/biharmonic lattice Green's functions (Theorem 2.1), represents homogeneous fields by sources on thin staggered boundary layers, enforces Dirichlet data by local cut-point interpolation, removes hydrostatic null modes by sampled-normal rank updates, and accelerates/solves the resulting dense boundary system with FFT convolutions and a componentwise discrete Calderón preconditioner built from the scalar Laplace LGF. Numerical experiments cover interior, multiply connected, narrow-gap, body-forced, corner-eddy, and exterior configurations, reporting roughly second-order velocity convergence, small discrete divergence, and reduced condition numbers.

Significance. If the construction is fully rigorous, the paper offers a valuable new discretize-then-represent boundary algebraic framework for Stokes flow that avoids singular quadrature and artificial outer boundary conditions, with an explicit, reusable lattice Green's function pair and a promising preconditioning strategy. The derivation of the MAC Stokes LGF is clean and self-contained, and the numerical evidence is extensive and consistently supportive: second-order velocity convergence across many geometries, divergence at solver accuracy, and substantial condition-number reduction. The main value lies in the combination of staggered-grid discrete potential theory with unfitted boundary closure. However, two analytical pillars of the method are asserted rather than proved: the sampled-normal rank completion that removes the hydrostatic nullspace, and the discrete Calderón identity underlying the preconditioner. These gaps currently limit the strength of the general algorithmic claims, even though the numerical results are encouraging.

major comments (3)
  1. [Section 4, Eq. (4.17), Remark 4.1] The rank completion is the keystone of the boundary closure, but it is asserted rather than proved. Theorem 3.2 only identifies ker S−; it says nothing about ker A = Φ+S+ + Φ−S−, and Remark 4.1 gives no argument that the sampled normal nΓ is aligned with the algebraic hydrostatic mode zh = ∇hχ|γ− for arbitrary cut configurations. For K objects one additionally needs NΓᵀZ to be nonsingular. If nΓ misses a null direction (e.g., a boundary almost tangent to a grid line, or a non-convex shape with multiple intersections on one grid line), Ac in (4.17)/(4.19) remains singular and the solve is ill-posed. The experiments in Section 9 show the completion worked for the tested geometries but do not establish the general claim. Please add (i) a direct numerical check of the smallest singular value/algebraic nullspace of A and Ac for a range of unfitted geometries, including near-tangent cuts and m
  2. [Section 3.4, Lemma 3.1 and Theorem 3.2] The energy argument assumes that the regularized LGF potential generated by a force-balanced density has finite discrete energy and that the only finite-energy homogeneous discrete Stokes field with zero trace on γ− is zero. For 2D free-space Stokes, a force dipole has velocity decaying as 1/r, so the continuous energy integral is logarithmically divergent; whether the chosen kernel regularization/gauge restores finite energy on the lattice is not shown. Moreover the boundary system is not actually constrained to the force-balanced class during the solve, so the relevant object is the nullspace of A, which includes the interpolation operators Φ±; non-hydrostatic null vectors introduced by the closure are not ruled out. Please prove or carefully state these spectral assumptions, and provide a numerical nullspace diagnostic for A.
  3. [Section 5, Eq. (5.6)] The discrete Calderón identity Vα(Jα+Wα) = −Kα(Iα+Kα) is stated as a fact for the unfitted staggered boundary layer with incidence scaling Jα. The cited source [29] establishes related identities for fitted lattice boundaries; the unfitted case with variable incidence is new and no proof is supplied. Since the identity underlies the claimed mesh-independent preconditioning, the paper should either prove it under explicit assumptions or present it as a heuristic supported by the numerical conditioning data. At minimum, state the precise sense in which the identity holds (exactly, up to O(h), or only in a spectral average sense).
minor comments (4)
  1. [Abstract and Section 9.3] The abstract states 'second-order velocity and pressure convergence', but Tables 3 and 11 show pressure rates around 1.55–1.80 in the Taylor–Couette and single-obstacle exterior cases. I suggest qualifying the claim as 'approximately second-order velocity and close to second-order pressure in most configurations'.
  2. [Section 9.1.5] The Moffatt-eddy ratios are compared to the asymptotic value 2.10 on a single grid. A brief refinement study of the eddy-center locations would substantiate the claim that the method resolves the asymptotic scaling rather than producing the right value by coincidence.
  3. [Section 9.1.2, Figure 7] The text says the total runtime scaling is 'consistent with near-optimal O(N² log N)', but the reported total times approximately quadruple per refinement, which is also consistent with O(N²). The paper should state which component dominates the log factor or present a more detailed scaling breakdown.
  4. [Section 7] For body forcing that is not compactly supported (e.g., the trigonometric forcing in Section 9.1.6), the FFT-based convolution truncates the infinite source distribution after padding. The paper should state the implicit decay/truncation assumption and note that sufficient padding is needed to control aliasing for non-decaying data.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the Stokes kernel construction, boundary algebraic equation, and preconditioner are independently derived and validated against external benchmarks.

full rationale

The derivation chain is self-contained. Theorem 2.1 constructs the Stokes velocity and pressure lattice Green's functions from the scalar Laplace and biharmonic LGFs via staggered Fourier transforms; the target property (solving the discrete Stokes equations) is proven, not assumed. The boundary algebraic equation arises from the kernel's translation invariance and the layer-potential representation, with no fitted parameter renamed as a prediction. The rank-one completion in Eq. (4.17) is asserted rather than proved — Remark 4.1 claims nΓ is the cut-point representation of the exact discrete null vector 'up to discretization and scaling' without proof; this is a correctness risk (the reviewer's skeptical point is legitimate) but it is not circular, because the normal vector is geometric and is not fitted to the convergence rates, condition numbers, or Moffatt ratios reported later. The numerical claims are compared to exact manufactured solutions, the analytical Taylor–Couette profile, and the independent Moffatt theoretical ratio 2.10. Self-citations ([16], [17]) provide the local-boundary-interpolation scaffolding, but the Stokes-specific contributions — the coupled kernel pair, discrete incompressibility preservation, hydrostatic rank completion, and componentwise Calderón preconditioner — are independently constructed and externally validated. No load-bearing argument reduces to a self-citation chain or to a fitted input.

Assumptions & free parameters 2 free parameters · 4 assumptions · 1 invented entities

The method rests on standard lattice Green's function and discrete potential theory plus two paper-specific constructs: the sampled-normal rank completion and the componentwise Calderón preconditioner. The latter's identity is imported from a dissertation and the former is unproved in the general cut-cell case.

free parameters (2)
  • tau (rank-completion shift) = τ = ||A||_F / sqrt(n)
    Introduced in Remark 4.1 to lift the hydrostatic nullspace; magnitude chosen heuristically as mean singular-value scale. The solvability of the completed system and its conditioning depend on this choice; no error analysis is given.
  • c_alpha (preconditioner deflation shift) = c_α = 0.25 / (e_α^T V_α e_α)
    Chosen in eq (5.8) to map the constant mode of the hyper-singular operator to the spectral cluster center 0.25. It is computed from the boundary-layer operator, not from data, but it is a tuning parameter of the regularized preconditioner.
assumptions (4)
  • standard math Duffin–Shelly identity (2.28) gives a compatible regularization of the biharmonic lattice Green's function H from the regularized Laplace LGF G
    Used in Section 2 to construct H; cited from [27]. Accepted as background result.
  • ad hoc to paper The discrete Calderón identity Vα(Jα+Wα) = -Kα(Iα+Kα) holds for the unfitted staggered boundary layer with incidence scaling Jα
    Section 5, eq (5.6); adapted from [29] without derivation in this paper. The spectral reasoning for the preconditioner depends on this identity.
  • domain assumption The only finite-energy homogeneous discrete Stokes velocity field on the infinite lattice with zero trace on γ− is zero
    Theorem 3.2 proof; used to show ker S− = Zh. Not proven in the paper; it is a plausible uniqueness property for the discrete exterior problem.
  • ad hoc to paper The sampled normal vector nΓ is a valid discrete proxy for the hydrostatic null vector zh = ∇hχ|γ− up to discretization and scaling
    Remark 4.1 after eq (4.17); no proof. Essential for the rank-one completion to actually remove the hydrostatic nullspace.
invented entities (1)
  • No new physical entities
    purpose: The paper introduces only discrete algebraic constructs (LGFs, boundary layers, preconditioners).
    There is no new particle, force, or physical object; all novelties are mathematical/numerical objects internal to the method.

how reviews work

0 comments
Cite this review

Pith. "Pith review of An unfitted boundary algebraic equation method with Calder\'on preconditioning for 2D Stokes flow in irregular geometry." pith.science (2026). https://pith.science/paper/JMB6DHNU

@misc{pith2026260721295,
  author       = {Pith},
  title        = {Pith review of: An unfitted boundary algebraic equation method with Calder\'on preconditioning for 2D Stokes flow in irregular geometry},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JMB6DHNU}},
  note         = {Machine review of arXiv:2607.21295}
}
read the original abstract

We present an unfitted boundary algebraic equation method for the two-dimensional exterior/interior Stokes equations on a staggered MAC grid. By constructing an explicit free-space pair of velocity and pressure lattice Green's functions (LGFs) from free-space Laplace LGFs, we represent homogeneous fields using sources supported exclusively on thin staggered boundary layers. This formulation imposes physical Dirichlet data at cut points via local interpolation, while sampled-normal rank updates remove hydrostatic null modes associated with single or multiple obstacles. The workflow parallels that of classical boundary integral formulations and requires no artificial boundary conditions for exterior flows, but follows a discretize-then-represent route and does not require singular/near-singular quadrature. The resulting dense boundary system is solved via GMRES, utilizing a componentwise discrete Calder\'on preconditioner built from the scalar Laplace kernel and padded FFTs for fast volume convolutions. Extensive numerical validation, including multiply connected domains, narrow gaps, and Moffatt eddies, confirms discrete incompressibility to solver accuracy and recovers the expected Moffatt eddy scaling. We achieve second-order velocity and pressure convergence and bound maximum discrete divergence within numerical accuracy. The discrete Calder\'on preconditioner reduces the condition number by orders of magnitude and yields nearly mesh-independent conditioning in exterior configurations, while remaining effective---though more demanding---for narrow-gap and fine-grid interior problems.

Figures

Figures reproduced from arXiv: 2607.21295 by the authors.

Figure 1
Figure 1. MAC scheme The horizontal velocity u is stored at vertical cell faces, um1+1/2,m2 ≈ u(xm1+1/2 , ym2 ), (2.7) the vertical velocity v is stored at horizontal cell faces, vm1,m2+1/2 ≈ v(xm1 , ym2+1/2 ), (2.8) and the pressure is stored at cell centers, pm1,m2 ≈ p(xm1 , ym2 ). (2.9) 4 [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Illustration of γ u ± and γ v ± (solid triangles and squares denote γ u − and γ v − and hollow triangles and squares denote γ u + and γ v +) 3.3 Boundary algebraic trace equations Theorem 3.1 pairs q u and q v on the two staggered faces of a common cell. Near an unfitted boundary, however, |γ u | and |γ v | need not agree. The next result therefore allows the two densities to be supported independently on their resp… view at source ↗
Figure 3
Figure 3. Cut-cell interpolation for staggered velocity components. (Contribution of solid [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Maximum-norm convergence for ellipses with different aspect ratios [PITH_FULL_IMAGE:figures/full_fig_p021_4.png]
Figure 5
Figure 5. Figure 5: Velocity magnitude with streamlines (top) and pressure (bottom) for ellipse aspect [PITH_FULL_IMAGE:figures/full_fig_p021_5.png]
Figure 6
Figure 6. Figure 6: Taylor–Couette solution at N = 1023: velocity field, computed azimuthal profile, and pointwise azimuthal-velocity error. N Eu Rate Ev Rate Ep Rate 127 5.2835 × 10−5 – 5.2835 × 10−5 – 1.2282 × 10−3 – 255 1.3496 × 10−5 1.97 1.3496 × 10−5 1.97 4.1894 × 10−4 1.55 511 3.433…
Figure 7
Figure 7. Figure 7: GMRES relative residuals and runtimes for the Taylor–Couette benchmark. [PITH_FULL_IMAGE:figures/full_fig_p023_7.png]
Figure 8
Figure 8. Figure 8: Velocity magnitude with streamlines and pressure in the two-hole domain at [PITH_FULL_IMAGE:figures/full_fig_p025_8.png]
Figure 9
Figure 9. Figure 9: Velocity magnitude with streamlines and pressure in the narrow-gap domain at [PITH_FULL_IMAGE:figures/full_fig_p026_9.png]
Figure 10
Figure 10. Figure 10: Moffatt eddies in a 30◦ wedge at N = 1023; color denotes log10 |u|. 9.1.6 Interior flow with body forcing To verify the volume-potential construction of Section 6, we return to the unit disk (α = 1 in the ellipse study) and prescribe the smooth exact fields u = sin(πx…
Figure 11
Figure 11. Figure 11: Velocity magnitude with streamlines and pressure for the single-obstacle exterior flow [PITH_FULL_IMAGE:figures/full_fig_p029_11.png]
Figure 12
Figure 12. Figure 12: Velocity magnitude with streamlines and pressure for the three-obstacle exterior flow [PITH_FULL_IMAGE:figures/full_fig_p030_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 11 canonical work pages

  1. [29]

    Vikram Bhamidipati.Forward and Inverse Modeling of Conducting Lattices Using Lattice Green’s Functions. Ph.d. dissertation, The University of Texas at Austin, Austin, TX, December 2021

    2021

  2. [1]

    The immersed boundary method.Acta Numerica, 11:479–517, 2002

    Charles S Peskin. The immersed boundary method.Acta Numerica, 11:479–517, 2002. doi: 10.1017/S0962492902000077

    doi:10.1017/s0962492902000077 2002

  3. [2]

    The immersed interface method for elliptic equations with discontinuous coefficients and singular sources.SIAM Journal on Numerical Analysis, 31(4):1019–1044, 1994

    Randall J LeVeque and Zhilin Li. The immersed interface method for elliptic equations with discontinuous coefficients and singular sources.SIAM Journal on Numerical Analysis, 31(4):1019–1044, 1994. doi: 10.1137/0731054

    doi:10.1137/0731054 1994

  4. [3]

    A fast immersed interface method for solving Stokes flows on irregular domains.Computers & Fluids, 38(10):1973–1983, 2009

    Zhijun Tan, KM Lim, and BC Khoo. A fast immersed interface method for solving Stokes flows on irregular domains.Computers & Fluids, 38(10):1973–1983, 2009. doi: 10.1016/j.compfluid.2009.06.004

    work page doi:10.1016/j.compfluid.2009.06.004 1973

  5. [4]

    A fictitious domain method for Dirichlet problem and applications.Computer Methods in Applied Mechanics and Engineering, 111(3-4):283–303, 1994

    Roland Glowinski, Tsorng-Whay Pan, and Jacques Periaux. A fictitious domain method for Dirichlet problem and applications.Computer Methods in Applied Mechanics and Engineering, 111(3-4):283–303, 1994. doi: 10.1016/0045-7825(94)90135-X

    work page doi:10.1016/0045-7825(94)90135-x 1994

  6. [5]

    Fictitious domain methods using cut elements: III

    Erik Burman and Peter Hansbo. Fictitious domain methods using cut elements: III. a stabilized Nitsche method for Stokes’ problem.ESAIM: Mathematical Modelling and Numerical Analysis, 48(3):859–874, 2014. doi: 10.1051/m2an/2013123

    arXiv 2014

  7. [6]

    A cut finite element method for a Stokes interface problem.Applied Numerical Mathematics, 85:90–114, 2014

    Peter Hansbo, Mats G Larson, and Sara Zahedi. A cut finite element method for a Stokes interface problem.Applied Numerical Mathematics, 85:90–114, 2014. doi: 10.1016/j.apnum. 2014.06.009

    doi:10.1016/j.apnum 2014

  8. [7]

    Cambridge University Press, 1992

    Constantine Pozrikidis.Boundary integral and singularity methods for linearized viscous flow. Cambridge University Press, 1992. doi: 10.1017/CBO9780511624124

    work page doi:10.1017/cbo9780511624124 1992

Show all 32 references
  1. [8]

    A fast multipole method for the three- dimensional Stokes equations.Journal of Computational Physics, 227(3):1613–1619, 2008

    Anna-Karin Tornberg and Leslie Greengard. A fast multipole method for the three- dimensional Stokes equations.Journal of Computational Physics, 227(3):1613–1619, 2008. doi: 10.1016/j.jcp.2007.06.029

    2008 doi

  2. [9]

    A kernel-free boundary integral method for elliptic boundary value problems.Journal of Computational Physics, 227(2):1046–1074, 2007

    Wenjun Ying and Craig S Henriquez. A kernel-free boundary integral method for elliptic boundary value problems.Journal of Computational Physics, 227(2):1046–1074, 2007. doi: 10.1016/j.jcp.2007.08.021. 32

    2007 doi

  3. [10]

    A kernel-free boundary integral method for implicitly defined surfaces.Journal of Computational Physics, 252:606–624, 2013

    Wenjun Ying and Wei-Cheng Wang. A kernel-free boundary integral method for implicitly defined surfaces.Journal of Computational Physics, 252:606–624, 2013. doi: 10.1016/j.jcp. 2013.06.019

    2013 doi

  4. [11]

    A correction function-based kernel-free boundary integral method for elliptic PDEs with implicitly defined interfaces.Journal of Computational Physics, 496:112545, 2024

    Han Zhou and Wenjun Ying. A correction function-based kernel-free boundary integral method for elliptic PDEs with implicitly defined interfaces.Journal of Computational Physics, 496:112545, 2024. doi: 10.1016/j.jcp.2023.112545

    2024

  5. [12]

    A correction function-based KFBI method for Brinkman interface problems

    Han Zhou and Wenjun Ying. A correction function-based KFBI method for Brinkman interface problems. 2026. doi: 10.48550/arXiv.2604.15509

    2026 doi arXiv

  6. [13]

    Discrete potential theory.Duke Math

    Richard J Duffin. Discrete potential theory.Duke Math. J., 20(1):233–251, 1953. doi: 10.1215/S0012-7094-53-02023-7

    1953 doi

  7. [14]

    Boundary algebraic equations for lattice problems.Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 465(2108):2489–2503, 2009

    Per-Gunnar Martinsson and Gregory J Rodin. Boundary algebraic equations for lattice problems.Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 465(2108):2489–2503, 2009. doi: 10.1098/rspa.2008.0473

    2009

  8. [15]

    Ryaben’kii.Method of Difference Potentials and its applications, volume 30

    Viktor S. Ryaben’kii.Method of Difference Potentials and its applications, volume 30. Springer Science & Business Media, 2012. doi: 10.1007/978-3-642-56344-7

    2012 doi

  9. [16]

    Local-basis difference potentials method for elliptic PDEs in complex geometry

    Qing Xia. Local-basis difference potentials method for elliptic PDEs in complex geometry. Journal of Computational Physics, 488:112246, 2023. doi: 10.1016/j.jcp.2023.112246

    2023

  10. [17]

    A geometrically robust unfitted boundary algebraic equation method based on discrete potentials and local basis functions.Journal of Computational Physics, 561:114957,

    Qing Xia. A geometrically robust unfitted boundary algebraic equation method based on discrete potentials and local basis functions.Journal of Computational Physics, 561:114957,

  11. [18]

    Asymptotic expansions of lattice Green’s functions.Proceedings of the Royal Society of London

    Per-Gunnar Martinsson and Gregory J Rodin. Asymptotic expansions of lattice Green’s functions.Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences, 458(2027):2609–2622, 2002. doi: 10.1098/rspa.2002.0985

    2027

  12. [19]

    A fast solver for Poisson problems on infinite regular lattices.Journal of Computational and Applied Mathematics, 258:42–56,

    Adrianna Gillman and Per-Gunnar Martinsson. A fast solver for Poisson problems on infinite regular lattices.Journal of Computational and Applied Mathematics, 258:42–56,

  13. [20]

    Afastmulti-resolution lattice Green’s function method for elliptic difference equations.Journal of Computational Physics, 407:109270, 2020

    BenediktDorschner, KeYu, GianmarcoMengaldo, andTimColonius. Afastmulti-resolution lattice Green’s function method for elliptic difference equations.Journal of Computational Physics, 407:109270, 2020. doi: 10.1016/j.jcp.2020.109270

    2020

  14. [21]

    Hodapp, G

    M. Hodapp, G. Anciaux, and W. A. Curtin. Lattice Green function methods for atom- istic/continuum coupling: Theory and data-sparse implementation.Computer Methods in Applied Mechanics and Engineering, 348:1039–1075, 2019. doi: 10.1016/j.cma.2019.02.006

    2019 doi

  15. [22]

    Kallivokas, and Gregory J

    Vikram Bhamidipati, Loukas F. Kallivokas, and Gregory J. Rodin. The inverse problem for conducting defective lattices.Computer Methods in Applied Mechanics and Engineering, 393:114788, 2022. doi: 10.1016/j.cma.2022.114788

    2022

  16. [23]

    A fast lattice Green’s function method for solving viscous incompressible flows on unbounded domains.Journal of Computational Physics, 316:360–384, 2016

    Sebastian Liska and Tim Colonius. A fast lattice Green’s function method for solving viscous incompressible flows on unbounded domains.Journal of Computational Physics, 316:360–384, 2016. doi: 10.1016/j.jcp.2016.04.023

    2016 doi

  17. [24]

    A fast immersed boundary method for external incompressible viscous flows using lattice Green’s functions.Journal of Computational Physics, 331:257–279, 2017

    Sebastian Liska and Tim Colonius. A fast immersed boundary method for external incompressible viscous flows using lattice Green’s functions.Journal of Computational Physics, 331:257–279, 2017. doi: 10.1016/j.jcp.2016.11.034. 33

    2017 doi

  18. [25]

    Jeremy A. K. Horwitz, Gianluca Iaccarino, John K. Eaton, and Ali Mani. The discrete Green’s function paradigm for two-way coupled Euler–Lagrange simulation.Journal of Fluid Mechanics, 931:A3, 2022. doi: 10.1017/jfm.2021.928

    2022 doi

  19. [26]

    Borwein, M

    Jonathan M. Borwein, M. L. Glasser, R. C. McPhedran, J. G. Wan, and I. J. Zucker.Lattice sums then and now. Number 150 in Encyclopedia of Mathematics and its Applications. Cambridge University Press, 2013. doi: 10.1017/CBO9781139626804

    2013 doi

  20. [27]

    R. J. Duffin and E. P. Shelly. Difference equations of polyharmonic type.Duke Mathematical Journal, 25(2):209–238, June 1958. doi: 10.1215/S0012-7094-58-02521-3

    1958 doi

  21. [28]

    On Galerkin difference methods.Journal of Computational Physics, 313:310–327, 2016

    Jeffrey W Banks and Thomas Hagstrom. On Galerkin difference methods.Journal of Computational Physics, 313:310–327, 2016. doi: 10.1016/j.jcp.2016.02.042

    2016 doi

  22. [30]

    Viscous and resistive eddies near a sharp corner.Journal of Fluid Mechanics, 18(1):1–18, 1964

    H Keith Moffatt. Viscous and resistive eddies near a sharp corner.Journal of Fluid Mechanics, 18(1):1–18, 1964. doi: 10.1017/S0022112064000015. 34

    1964 doi

  23. [2014]

    doi: 10.1016/j.cam.2013.09.003

    2013 doi

  24. [2026]

    doi: 10.1016/j.jcp.2026.114957

    2026

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.