Pith. sign in

REVIEW 4 major objections 4 minor 33 references

Robust space-time multiscale upscaling via multicontinuum homogenization for evolving perforated media

T0 review · 4 major / 4 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 A space-time multicontinuum upscaling procedure reduces parabolic flow in evolving perforated domains to a coupled macroscopic system with computable coefficients.

desk verdict Useful space-time extension of multicontinuum homogenization, but the upscaled equation rests on an unproved term-dropping step and the numerics use ε=H, so the derivation gap is real. read the letter →

arxiv 2506.21104 v1 pith:NI4VKW2H submitted 2025-06-26 math.NA cs.NA

Wei Xie , Viet Ha Hoang , Yin Yang , Yunqing Huang This is my paper · ORCID
classification math.NAcs.NA MSC 35B2765M6065N3076S05
keywords multicontinuumhomogenizationspace-timeupscalingperforateddomainsevolvingmacroscopicequationsparabolicPDErepresentativevolumeelementporousmediaflow
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 sets out to show that a diffusion-like parabolic flow in a pore structure that shrinks over time, through clogging, deposition, or dissolution, can be captured by a small system of macroscopic equations without resolving the moving fine-scale geometry. The construction separates the domain into continua by channel width and writes the fine solution as a truncated expansion $u \approx \phi_i U_i + \phi_i^m \frac{\partial U_i}{\partial x_m}$, with basis functions computed once on oversampled representative volume elements that include time-derivative terms. Inserting this expansion into the weak form and dropping cross terms yields the coupled macroscopic system $D_{ji}\frac{\partial U_i}{\partial t}+B_{ji}U_i-\partial_n(B^{mn}_{ji}\partial_m U_i)=b_j$ with computable coefficients. A sympathetic reader would care because, if the upscaled model is right, large-scale simulations of reactive transport and pore clogging can run on a fixed coarse grid while still tracking the average behavior in each channel family.

What carries the argument

The carrying object is the two-term expansion $u \approx \phi_i U_i + \phi_i^m \partial_m U_i$ over a representative volume element $R_p$, with basis functions defined by the constrained local problems (8)-(9) on an oversampled region $R_p^+$. The constraints force $\phi_i$ to have unit mean over continuum $i$ and zero mean over the others, and force $\phi_i^m$ to match the mean of $x^m-c^m$ within continuum $i$; Lagrange multipliers $\beta$ and $\zeta$ enforce these averages, while the temporal terms in the cell problems transfer the domain evolution into the basis. The scaling estimates $\|\phi_i\|=O(1)$, $\|\nabla\phi_i\|=O(1/\varepsilon)$, $\|\phi_i^m\|=O(\varepsilon)$, $\|\nabla\phi_i^m\|=O(1)$ are the mechanism by which all unlisted cross terms are discarded, leaving the closed macroscopic system (14).

What would settle it

Take any one of the three test configurations and compute, at each time step and for each coarse block, the magnitudes of the dropped expressions $\int_{R_p}\phi_i\phi_j^n \partial_t U_i \partial_n V_j$ and $\int_{R_p}\kappa\nabla\phi_i\cdot\nabla\phi_j^n$ alongside the retained coefficients in (14). If the ratio of either class of dropped term to the corresponding retained term fails to stay small as $H\to 0$ and the domain contracts, the effective equations will not reproduce the per-continuum averages measured by (16), and the paper's central claim collapses.

Watch

Extended reading notes

Core claim

The central claim is that the truncated multicontinuum expansion, together with the constrained space-time cell problems (8)-(9), reduces the original parabolic fine-scale model (1) to the macroscopic system (14) plus its initial-condition counterpart (15). The coarse variables $U_i$ are the per-continuum spatial averages of the fine solution, the coefficient tensors $D_{ji}$, $B_{ji}$, $B^{mn}_{ji}$, and $b_j$ are assembled from the local basis functions, and the time-dependent cell problems make the basis functions aware of the moving geometry. In the three numerical settings reported, with a two-continuum channel geometry, permeability contrast of $10^2$, and prescribed contraction of the two channels at different rates, the multiscale averages stay close to the fine-grid per-continuum averages, with relative errors in the few-percent range for $H=1/10$ and smaller for $H=1/20$.

Load-bearing premise

The derivation moves from the upscaled variational problem to the strong form (14) by announcing that "other terms will be ignored by the symmetric properties or the scaling in (10)"; the whole coarse model rests on the unproved premise that every omitted cross term is genuinely negligible relative to the terms kept in (14).

Editorial extensions

If this is right

  • Assembling $D_{ji}$, $B_{ji}$, $B^{mn}_{ji}$, and $b_j$ from the local cell problems gives a coarse model that runs on a fixed mesh even while the underlying pores shrink, so the cost of a time step no longer scales with the number of fine-grid cells that disappear.
  • The macroscopic unknowns $U_i$ are per-continuum averages, so quantities such as the mass remaining in thin low-permeability channels are available directly from the coarse solution without post-processing the fine geometry.
  • Although the paper implements two continua, the construction is stated for arbitrary characteristic functions $\psi_i$, so channel networks with several width classes inherit the same local problems and the same coefficient formulas.
  • The initial-condition system (15) is derived by the same averaging, which lets the coarse model be started from the fine initial data without a separate fine-scale solve.

Reading between the lines

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

  • Editorial extension: the paper does not estimate the dropped cross terms, but the natural check is to bound the ratios of $\int_{R_p}\phi_i\phi_j^n \partial_t U_i \partial_n V_j$ and $\int_{R_p}\kappa\nabla\phi_i\cdot\nabla\phi_j^n$ to the retained coefficients; if these ratios grow like $\varepsilon/H$ as the domain contracts, the next-order basis functions $\phi^{mn}$ would have to be retained r
  • Editorial extension: the geometry in the experiments is prescribed by a fixed contraction rate; a coupled extension would let the continuum indicator functions $\psi_i$ evolve with the fine-scale solution, and the same local problems should still supply the coarse coefficients as long as the evolution is slow relative to the cell-problem time step.
  • Editorial extension: the oversampling depth is chosen globally and empirically; an a-posteriori indicator measuring the magnitude of the omitted cross terms on each coarse block could make the oversampling adaptive and cheaper on benign blocks.
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

4 major / 4 minor

Summary. The paper proposes a space-time multicontinuum homogenization framework for parabolic equations posed on time-evolving perforated domains. The method introduces local cell problems on oversampled representative volume elements (RVEs) that include temporal derivatives and domain evolution, and derives a macroscopic system (14) for per-continuum averages. The framework is tested on three numerical examples with shrinking domains and heterogeneous permeability fields.

Significance. If the derivation is correct, the framework offers a computable coarse-scale model for a class of dynamic porous-media problems, and the numerical experiments provide non-circular evidence because they compare against an independent fine-scale reference solution. The main strengths are the conceptual novelty of combining multicontinuum ideas with time-evolving geometry and the breadth of the numerical tests. However, the central derivation of the effective equation contains an unproved negligibility step, the local cell problems are not fully specified, and the numerical setting does not satisfy the scale-separation assumption used in that derivation. These gaps prevent the paper from establishing its main claim as it stands.

major comments (4)
  1. [Section 3, transition to Eq. (14)] The derivation of the strong form (14) from the displayed upscaled variational problem drops a number of cross terms with the statement that they are ignored 'by the symmetric properties or the scaling in the (10).' These include the mass coupling terms ∫ φ_i φ_j^n ∂U_i/∂t ∂V_j/∂x_n and ∫ φ_i^m φ_j V_j ∂^2 U_i/∂t∂x_m, as well as the mixed stiffness terms ∫ κ∇φ_i·∇φ_j^n and ∫ κ∇φ_i^m·∇φ_j. No symmetry property of the basis functions that would make these integrals vanish is established, and the scaling (10) only implies relative size O(ε/H) (or O(ε) in the unscaled integrals) compared with the retained terms. Since (14) is the paper's central output, this is a load-bearing gap. The authors should either provide a quantitative estimate for all omitted terms, justify their disappearance through explicit orthogonality properties, or numerically quantify their contribution.
  2. [Section 3, Eqs. (8)-(9)] The local cell problems (8)-(9) are not fully specified. No boundary condition on ∂R_p^+ is stated, and no initial condition for φ_i or φ_i^m is given beyond an elliptic equation at t=0. Without boundary conditions, the constrained variational problems are not well-posed: the bilinear form a(·,·) is not coercive on a space of functions without boundary constraints, and the constraints fix only averages. Since every effective coefficient in (14) is an integral of these basis functions, the well-posedness of (8)-(9) is a prerequisite for the method. The authors should specify the boundary condition on the oversampling boundary and provide a well-posedness argument, or at least state the discrete setting used in the numerical implementation.
  3. [Section 4, first paragraph] The numerical experiments set 'each coarse block is treated as a RVE,' so the RVE scale equals the coarse scale H. This means the parameter ε in the scaling (10) coincides with the coarse-grid size, and the O(ε/H) negligibility premise in the derivation of (14) is not satisfied (indeed ε/H = 1). The reported relative L2 errors, which are 2-10%, therefore do not support the claim that (14) is the correct effective equation; they show only that some coarse model reproduces the continuum averages in the tested cases. To validate the derivation, the experiments should either use an RVE that is strictly smaller than the coarse block, or the derivation should be extended to the case ε = H with an explicit justification of the dropped terms.
  4. [Section 3, Eqs. (12)-(13)] The approximation in (12) is written as ∫ ∂(φ_i U_i)/∂t v + ∫ ∂_m(φ_i^m U_i)/∂t v ≈ U_i ∫ ∂φ_i/∂t v + ∂_m U_i ∫ ∂φ_i^m/∂t v, but this omits the time-derivative terms ∂U_i/∂t ∫ φ_i v and ∂(∂_m U_i)/∂t ∫ φ_i^m v that arise from the product rule. The subsequent upscaled variational problem does contain a term V_j ∂U_i/∂t ∫ φ_i φ_j, so the display is inconsistent with the proceeding derivation. This notational inconsistency obscures which terms are being kept and which are being discarded, and it should be corrected.
minor comments (4)
  1. [Section 3, paragraph on scales] The displayed hierarchy 'h < ε < ε < H' uses the same symbol ε for two different scales (the heterogeneity scale and the RVE scale). Distinct symbols should be used to avoid confusion.
  2. [Section 3, Eq. (15)] The coefficient D^m_ji is defined, but the equation uses D^mn_ji. The index structure is inconsistent and should be aligned.
  3. [Section 4, Figures 7 and 8] The captions of Figures 7 and 8 state 'Example 1' but the corresponding text is for Example 3; the captions should be corrected.
  4. [Section 4, Tables 1-3] The text says the errors 'confirming the convergence' of the method, but only two coarse grid sizes are tested and no convergence rate is reported. A statement about the trend would be more cautious than a claim of confirmed convergence.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the multiscale model is benchmarked against an independent fine-scale solution, and no fitted parameter is renamed as a prediction.

full rationale

The paper's derivation chain is self-contained in the sense required for a circularity finding. The multicontinuum expansion (7) and the constrained local problems (8)-(9) define the basis functions and the macroscopic variables; the effective coefficients D_ji, B_ji, B^mn_ji and right-hand sides are computed from those local solutions and forcing data. The coarse model (14) is then solved and its per-continuum averages are compared in (16) with a fine-scale finite element reference solution of (1). No quantity entering the coarse model is fitted to the target averages, and the fine-scale solution is not used to construct the macro coefficients, so the numerical comparison is not a self-fulfilling check. The self-citation to [29] for 'Following our previous work' introduces the multicontinuum ansatz, but the ansatz is stated explicitly in the present paper and the space-time cell problems with temporal derivatives are new here; the scaling estimate (10) is attributed to [14], which is not a solely self-referential source. The main weakness noted by the reader - the sentence 'Here, other terms will be ignored by the symmetric properties or the scaling in the (10)' - is an unjustified approximation in the transition from the upscaled variational problem to the strong form (14), and the Section 4 setting with each coarse block as an RVE weakens the eps < H premise. However, this is a correctness and rigor limitation rather than circular reasoning: the omitted cross terms are not equivalent to the retained terms by construction, and no prediction reduces to its input. Accordingly, the appropriate score is 0.

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

The paper introduces no fitted constants and no new physical entities. Its central claim rests on five unproved modeling and well-posedness assumptions, most importantly the negligibility of the dropped cross terms in the macroscopic equation and the well-posedness of the time-dependent cell problems.

assumptions (5)
  • domain assumption The multiscale expansion (6) can be truncated after the first two terms and U_i varies slowly inside each RVE.
    Invoked around Eqs. (7) and after (11); no error estimate is given for the neglected second-order correctors.
  • ad hoc to paper The local cell problems (8)-(9) are well-posed on the time-evolving oversampled domain with the stated constraints.
    A parabolic problem is posed on (0,T] without an explicit initial value for φ_i or a boundary condition on ∂R_p^+; the existence argument is absent.
  • domain assumption The scaling estimates (10), established in [14] for static problems, remain valid for cell problems with time derivatives and shrinking domains.
    The relations ∥φ^m_i∥=O(ε), ∥∇φ^m_i∥=O(1) are cited from [14] and used to discard terms in the macroscopic equation.
  • domain assumption The 'standard localization argument' converts the sum over RVEs into a single coarse equation on the fixed domain Ω.
    Invoked before Eq. (11); no proof is supplied for a moving perforated geometry with oversampling.
  • ad hoc to paper All terms omitted from the strong form (14) are negligible by symmetry or by the scaling (10).
    This is the key unproved selection step; the paper states 'other terms will be ignored' without estimates.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Robust space-time multiscale upscaling via multicontinuum homogenization for evolving perforated media." pith.science (2026). https://pith.science/paper/NI4VKW2H

@misc{pith2026250621104,
  author       = {Pith},
  title        = {Pith review of: Robust space-time multiscale upscaling via multicontinuum homogenization for evolving perforated media},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NI4VKW2H}},
  note         = {Machine review of arXiv:2506.21104}
}
read the original abstract

Time-evolving perforated domains arise in many engineering and geoscientific applications, including reactive transport, particle deposition, and structural degradation in porous media. Accurately capturing the macroscopic behavior of such systems poses significant computational challenges due to the dynamic fine-scale geometries. In this paper, we develop a robust and generalizable multiscale modeling framework based on multicontinuum homogenization to derive effective macroscopic equations in shrinking domains. The method distinguishes multiple continua according to the physical characteristics (e.g., channel widths), and couples them via space-time local cell problems formulated on representative volume elements. These local problems incorporate temporal derivatives and domain evolution, ensuring consistency with underlying fine-scale dynamics. The resulting upscaled system yields computable macroscopic coefficients and is suitable for large-scale simulations. Several numerical experiments are presented to validate the accuracy, efficiency, and potential applicability of the method to complex time-dependent engineering problems.

Figures

Figures reproduced from arXiv: 2506.21104 by the authors.

Figure 1
Figure 1. Illustration of shrinking domain. In this work, we present a novel space-time multicontinuum homogenization framework tai￾lored for time-dependent problems in shrinking perforated domains. These domains evolve due to processes such as pore clogging or crack propagation. As a representative example, we consider 2 [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Illustration of the representative volume element (RVE) [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Fine-grid solutions at different time steps in Example 1. [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Comparison of average solutions at final time in Example 1. [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: Fine-grid solutions at different time steps in Example 2. [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Comparison of average solutions at final time in Example 2. [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Fine-grid solutions at different time steps in Example 1. [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]
Figure 8
Figure 8. Figure 8: Comparison of average solutions at final time in Example 1. [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 32 canonical work pages

  1. [1]

    Abdulle, E

    A. Abdulle, E. Weinan, B. Engquist, and E. Vanden-Eijnden. The heterogeneous multiscale method.Acta Numerica, 21:1–87, 2012

    work page 2012

  2. [2]

    Alikhanov, H

    A. Alikhanov, H. Bai, J. Huang, A. Tyrylgin, and Y. Yang. Multiscale model reduction for the time fractional thermoporoelasticity problem in fractured and heterogeneous media.Journal of Computational and Applied Mathematics, 455:116157, 2025

    work page 2025

  3. [3]

    Altmann, P

    R. Altmann, P. Henning, and D. Peterseim. Numerical homogenization beyond scale separa- tion.Acta Numerica, 30:1–86, 2021

    work page 2021

  4. [4]

    H. Bai, D. Ammosov, Y. Yang, W. Xie, and M. A. Kobaisi. Multicontinuum modeling of time- fractional diffusion-wave equation in heterogeneous media.arXiv preprint arXiv:2502.09428, 2025

    work page Pith review arXiv 2025

  5. [5]

    Bringedal, L

    C. Bringedal, L. Von Wolff, and I. S. Pop. Phase field modeling of precipitation and dissolution processes in porous media: Upscaling and numerical experiments.Multiscale Modeling & Simulation, 18(2):1076–1112, 2020

    work page 2020

  6. [6]

    Chung, Y

    E. Chung, Y. Efendiev, J. Galvis, and W. T. Leung. Multicontinuum homogenization. general theory and applications.Journal of Computational Physics, 510:112980, 2024

    work page 2024

  7. [7]

    Chung, Y

    E. Chung, Y. Efendiev, W. T. Leung, and S. Ye. Generalized multiscale finite element methods for space–time heterogeneous parabolic equations.Computers & Mathematics with Applica- tions, 76(2):419–437, 2018

    work page 2018

  8. [8]

    Chung, Y

    E. Chung, Y. Efendiev, and Y. Li. Space-time gmsfem for transport equations.GEM- International Journal on Geomathematics, 9(2):265–292, 2018

    work page 2018

Show all 33 references
  1. [9]

    Chung, J

    E. Chung, J. Hu, and S.-M. Pun. Convergence of the cem-gmsfem for stokes flows in hetero- geneous perforated domains.Journal of Computational and Applied Mathematics, 389:113327, 2021

    2021

  2. [10]

    E. T. Chung, Y. Efendiev, and W. T. Leung. Constraint energy minimizing generalized mul- tiscale finite element method.Computer Methods in Applied Mechanics and Engineering, 339:298–319, 2018. 15

    2018

  3. [11]

    E. T. Chung, Y. Efendiev, W. T. Leung, M. Vasilyeva, and Y. Wang. Non-local multi-continua upscaling for flows in heterogeneous fractured media.Journal of Computational Physics, 372:22–34, 2018

    2018

  4. [12]

    V. M. Devigne, I. S. Pop, C. Van Duijn, and T. Clopeau. A numerical scheme for the pore- scale simulation of crystal dissolution and precipitation in porous media.SIAM Journal on Numerical Analysis, 46(2):895–919, 2008

    2008

  5. [13]

    Efendiev, J

    Y. Efendiev, J. Galvis, and T. Y. Hou. Generalized multiscale finite element methods (gmsfem). Journal of Computational Physics, 251:116–135, 2013

    2013

  6. [14]

    Efendiev and W

    Y. Efendiev and W. T. Leung. Multicontinuum homogenization and its relation to nonlocal multicontinuum theories.Journal of Computational Physics, 474:111761, 2023

    2023

  7. [15]

    Efendiev, W

    Y. Efendiev, W. T. Leung, B. Shan, and M. Wang. Multicontinuum splitting scheme for multiscale flow problems.arXiv preprint arXiv:2410.05253, 2024

    2024 arXiv

  8. [16]

    Efendiev and A

    Y. Efendiev and A. Pankov. Numerical homogenization of nonlinear random parabolic opera- tors.Multiscale Modeling & Simulation, 2(2):237–268, 2004

    2004

  9. [17]

    S. Fu, E. Chung, and G. Li. Edge multiscale methods for elliptic problems with heterogeneous coefficients.Journal of Computational Physics, 396:228–242, 2019

    2019

  10. [18]

    S. Fu, G. Li, R. Craster, and S. Guenneau. Wavelet-based edge multiscale finite element method for helmholtz problems in perforated domains.Multiscale Modeling & Simulation, 19(4):1684–1709, 2021

    2021

  11. [19]

    T. Y. Hou and X.-H. Wu. A multiscale finite element method for elliptic problems in composite materials and porous media.Journal of Computational Physics, 134(1):169–189, 1997

    1997

  12. [20]

    J. Hu, W. T. Leung, and E. Chung. Space–time non-local multi-continua multiscale method for channelized-media parabolic equations.Journal of Computational and Applied Mathematics, page 116669, 2025

    2025

  13. [21]

    V. V. Jikov, S. M. Kozlov, and O. A. Oleinik.Homogenization of differential operators and integral functionals. Springer Science & Business Media, 2012

    2012

  14. [22]

    Le Bris, F

    C. Le Bris, F. Legoll, and A. Lozinski. An msfem type approach for perforated domains. Multiscale Modeling & Simulation, 12(3):1046–1077, 2014

    2014

  15. [23]

    W. T. Leung. Some convergence analysis for multicontinuum homogenization.arXiv preprint arXiv:2401.12799, 2024

    2024 arXiv

  16. [24]

    W. T. Leung, E. T. Chung, Y. Efendiev, M. Vasilyeva, and M. Wheeler. Space-time nonlin- ear upscaling framework using nonlocal multicontinuum approach.International Journal for Multiscale Computational Engineering, 17(5), 2019

    2019

  17. [25]

    Miehe and S

    C. Miehe and S. Mauthe. Phase field modeling of fracture in multi-physics problems. part iii. crack driving forces in hydro-poro-elasticity and hydraulic fracturing of fluid-saturated porous media.Computer Methods in Applied Mechanics and Engineering, 304:619–655, 2016. 16

    2016

  18. [26]

    P. Ming, P. Zhang, et al. Analysis of the heterogeneous multiscale method for elliptic homog- enization problems.Journal of the American Mathematical Society, 18(1):121–156, 2005

    2005

  19. [27]

    G. A. Pavliotis and A. Stuart.Multiscale methods: averaging and homogenization, volume 53. Springer Science & Business Media, 2008

    2008

  20. [28]

    S. Wang, H. Li, R. Wang, R. Tian, Q. Sun, and Y. Ma. Numerical simulation of flow behavior of particles in a porous media based on cfd-dem.Journal of Petroleum Science and Engineering, 171:140–152, 2018

    2018

  21. [29]

    W. Xie, Y. Efendiev, Y. Huang, W. T. Leung, and Y. Yang. Multicontinuum homogenization in perforated domains.Journal of Computational Physics, 530:113845, 2025

    2025

  22. [30]

    W. Xie, J. Galvis, Y. Yang, and Y. Huang. Multiscale modeling of wave propagation with exponential integration and gmsfem.Communications in Nonlinear Science and Numerical Simulation, page 108825, 2025

    2025

  23. [31]

    W. Xie, V. H. Hoang, Y. Yang, and Y. Huang. A hierarchical approach for multicontinuum homogenization in high contrast media.arXiv preprint arXiv:2503.01276, 2025

    2025 arXiv

  24. [32]

    W. Xie, Y. Yang, E. Chung, and Y. Huang. Cem-gmsfem for poisson equations in heterogeneous perforated domains.Multiscale Modeling & Simulation, 22(4):1683–1708, 2024

    2024

  25. [33]

    Ye and E

    C. Ye and E. T. Chung. Constraint energy minimizing generalized multiscale finite element method for inhomogeneous boundary value problems with high contrast coefficients.Multiscale Modeling & Simulation, 21(1):194–217, 2023. 17

    2023

Pith tools

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