Pith. sign in

REVIEW 4 major objections 5 minor 39 references

Multicontinuum splitting schemes for multiscale wave problems

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper constructs two partially explicit time-stepping schemes for the high-contrast wave equation that split fast and slow dynamics so that the stability condition is claimed to be independent of the coefficient contrast.

desk verdict A useful, mostly correct scheme paper whose headline claim of contrast-independent stability is deferred to an unpublished preprint and not closed by the numerics. read the letter →

arxiv 2506.01670 v1 pith:RR4RZRXC submitted 2025-06-02 math.NA cs.NA

Mohsen Alshahrani , Buzheng Shan This is my paper
classification math.NAcs.NA MSC 65M1265M6035L05
keywords multicontinuumhomogenizationpartiallyexplicittimediscretizationhigh-contrastwaveequationdiscreteenergyconservationcontrast-independentstabilitymultiscalebasisfunctionsgeneralizedeigenvaluedecompositionpropagationinheterogeneousmedia
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

Multicontinuum homogenization represents the solution of the high-contrast wave equation through local averages on several continua, and the paper's central move is to split those continua into a fast, high-value group and a slow, low-value group. On top of that split the authors build two three-layer time integration schemes that solve the fast group implicitly and the slow group explicitly. The main claim is that the resulting stability conditions, (3.26) and (3.29), bound the time step only by the slow group's energy norms, and that with properly chosen continua those norms, and hence the time step, are independent of the coefficient contrast. The numerical examples show the schemes matching the accuracy of the fully implicit discretization while the ordinary explicit scheme diverges with the same step size; the practical stake is wave simulation in high-contrast media at explicit-like cost without contrast-driven time-step restrictions.

What carries the argument

The machinery is the direct-sum decomposition $V_{\mathrm{mc},H}=V_{\mathrm{mc},1,H}\oplus V_{\mathrm{mc},2,H}$ of the discrete multicontinuum space, together with the bilinear forms $m_{ij}$, $a_{ij}$, and $c_{ij}$ defined through downscaling operators $T_j$. Each continuum is represented by multiscale basis functions obtained from local energy-minimizing cell problems with constraints on the averages of the solution; choosing indices $I_1$ for fast/high-value continua and $I_2$ for slow/low-value continua puts all contrast in the implicitly treated block. The stability proof runs through an exactly conserved discrete energy $E^{n+1/2}$; a strengthened Cauchy-Schwarz constant $\gamma$ controls cross-coupling, and the final conditions bound $\tau^2$ by $2(1-\gamma^2)$ times the $m_{22}$-to-$a_{22}$ (or $m_{22}$-to-$(a_{22}+c_{22})$) ratio of the slow subspace, which is exactly the quantity the cell-problem construction is claimed to make contrast-independent.

What would settle it

In the layered two-continuum example (5.3), compute the stability ratio $\inf_W \|W\|_{m_{22}}^2 / \|W\|_{a_{22}}^2$ for contrast values $10$, $10^3$, and $10^6$ while keeping the coarse mesh fixed; if this ratio shrinks as the contrast grows, the stability condition is contrast-dependent and the central claim fails.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that the standard multicontinuum homogenization expansion for the wave equation can be organized into two subspaces, one carrying the fast (high-contrast) dynamics and one the slow dynamics, so that the fully implicit semidiscrete problem can be replaced by the partially explicit three-layer schemes (3.8) and (3.9). The schemes conserve a discrete energy exactly (Theorem 1) and are stable under the conditions (3.26) and (3.29), which involve only the mass and stiffness norms of the explicitly treated slow subspace. Since, by the construction of the cell problems, those norms are claimed to be independent of the contrast for continua placed in low-value regions, the stability conditions are contrast-independent, and the method is presented as reaching essentially the accuracy of the fully implicit discretization at lower cost. The paper also proposes an optimized decomposition, through a tensor Rayleigh quotient problem and a cheaper generalized eigenvalue problem, to relax the stability restriction further.

Load-bearing premise

The headline result rests on an unproved assertion, deferred to a companion paper, that the cell-problem construction makes the energy stored in the low-coefficient continua insensitive to how large the high coefficient is.

Editorial extensions

If this is right

  • For two-value coefficients, choosing the high-value continuum as the implicit component and the low-value continuum as the explicit component makes the time-step restriction independent of the ratio $\kappa_{\max}/\kappa_{\min}$, so the same $\tau$ is expected to work as the contrast is raised.
  • The partially explicit schemes cost less per step than the fully implicit scheme because only the fast block requires an implicit solve, while the numerical tests show errors nearly identical to the implicit reference solution.
  • When the two subspaces are $L^2$-orthogonal, scheme 2 decouples, so the explicit part can be updated independently of the implicit solve; the paper notes that mass lumping can remove the remaining coupling in scheme 1.
  • The optimized decomposition based on a generalized eigenvalue problem selects continua by taking linear combinations of the original ones, and the resulting stability restriction scales roughly like $H\lambda_{i_0}^{-1/2}$, which relaxes the step-size bound without increasing computational cost.

Reading between the lines

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

  • If the contrast-independence proof transfers from the flow cell problems to the wave cell problems, the same splitting strategy should extend to other second-order hyperbolic systems, such as elastic waves or acoustics in fractured media, where high-contrast coefficients appear.
  • The tensor Rayleigh quotient formulation could be turned into an automatic continuum-selection tool for coefficient fields whose high- and low-value regions are not known in advance; the paper itself only treats predefined continua.
  • A natural testable extension is to push the contrast beyond $10^3$, the largest value reported here, and record whether the observed critical time step stays flat; under the paper's claim it should.
  • One could also test the optimized decomposition on coefficient fields with more than three continua, where the generalized eigenvalue problem would have to choose a genuinely mixed slow subspace rather than simply the low-value regions.
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 / 5 minor

Summary. The paper proposes two three-level partially explicit time discretization schemes, (3.8) and (3.9), for the high-contrast wave equation, built on a multicontinuum homogenization space decomposition. Fast continua are updated implicitly and slow continua explicitly. A discrete energy identity is proved for scheme 1 (Theorem 1), and stability is derived as condition (3.26) (Theorem 2); an analogous stability condition (3.29) for scheme 2 is stated as Theorem 3 without proof. The paper claims that, with appropriately chosen continua, the stability conditions are independent of the coefficient contrast, citing the cell-problem construction in the overlapping-author preprint [37]. An optimized decomposition via a tensor Rayleigh quotient and a simplified generalized eigenvalue problem is discussed in Section 4. Numerical examples for layered and point fields with two and three continua show that the proposed schemes match the accuracy of the fully implicit scheme for contrast 10^3.

Significance. If the contrast-independence assertion were proved, the schemes would provide a practically useful way to remove the high-contrast CFL restriction while retaining accuracy comparable to implicit methods. The self-contained algebraic derivations of Theorem 1 and Theorem 2 are a genuine contribution, as is the formulation of the stability conditions in terms of the explicitly treated subspace. However, the headline property is not established within the manuscript: it is deferred to an unpublished preprint, and the numerical section does not vary the contrast. The paper is therefore conditionally acceptable after the central gap is closed.

major comments (4)
  1. [Section 3, after Eq. (3.29)] The claim that the stability conditions (3.26) and (3.29) are contrast-independent is the central advertised result, but the only support is the sentence following Theorem 3 stating that the m22-, a22-, and c22-norms are 'all independent of the contrast ... thanks to the construction of the cell problems [37]'. No estimate of these norms as functions of κmax/κmin is given, and [37] is an overlapping-author preprint on flow problems that does not obviously cover the wave-specific c22 term entering (3.29). Moreover, the constant γ in (3.24) appears in both conditions through the factor (1−γ2), and no argument shows that γ is bounded away from 1 uniformly in the contrast. Please provide a proof, or at minimum a quantitative estimate, of contrast-independence for all quantities appearing in (3.26) and (3.29).
  2. [Theorem 3] Theorem 3 is stated without proof: the text says only that 'similar energy conservation can be established' for scheme 2. Since scheme 2 is one of the two proposed time discretizations and its stability condition (3.29) is used in the numerical section, the energy identity and the stability argument should be written out in full or explicitly reduced to the proof of Theorem 2 with the c22 terms accounted for.
  3. [Section 4] The optimized decomposition section defers its central assertions to [37]: 'It can be shown that γ = 0' and 'The same stability results hold under certain assumption' are not proved, and the localization assumption that 'the decompositions for different coarse blocks are similar' is stated without justification. Because Section 4 is presented as a way to relax the stability conditions and decouple the schemes, these claims need to be either proved in the manuscript or clearly stated as assumptions with supporting numerical evidence.
  4. [Section 5] The numerical experiments do not test the contrast-independence claim. All examples use the single contrast ratio 10^3 and the single time step τ = 10^-3, and no maximum stable time step is reported as a function of κmax/κmin. To support the headline property, please add experiments with varying contrast (for example 10^2, 10^4, 10^5) and report the largest stable τ for each case, or otherwise verify empirically that the stability conditions are contrast-independent.
minor comments (5)
  1. [Section 1] The sentence 'the variation in material properties ... necessities extremely fine spatial discretization' should read 'necessitates', and the sentence is grammatically incomplete.
  2. [Section 3] The first line, 'Throughout this work, We begin by summarizing...', contains a stray comma and an uppercase 'We' in the middle of the sentence.
  3. [Equation (4.4)] The definition of the Rayleigh quotient contains unmatched parentheses, and the contraction A : (v⊗w) is not defined explicitly.
  4. [Section 6] The sentence 'requiring less computationally effort' should be 'requiring less computational effort'.
  5. [Tables 1-4] The tables report eigenvalues and eigenvectors but not the resulting stability restriction τmax; reporting this value would allow readers to check conditions (3.26) and (3.29) directly.

Circularity Check

2 steps flagged · score 4.0 of 10

Contrast-independence of the stability conditions is asserted via an overlapping-author preprint [37], while the discrete-energy stability proofs are self-contained.

  1. self citation load bearing [Section 3, paragraph after Theorem 3 (conditions (3.26) and (3.29))]
    "The remaining question is how to determine the partition sets I1 and I2. For cases such as two-value fields, I1 and I2 can respectively be chosen as the index sets for the continua associated with the high- and low-value regions. It turns out that the m22-norm, the a22-norm and the c22-norm are all independent of the contrast, resulting in contrast-independent stability conditions, thanks to the construction of the cell problems [37]."

    The advertised central benefit is that conditions (3.26) and (3.29) are contrast-independent. This is not proved in the paper: no estimate of the m22-, a22-, or c22-norms as functions of the contrast is given, and the only support is the self-cited flow preprint [37]. Moreover, the c22 term appears specifically in the wave scheme 2 condition (3.29), but no argument shows that the flow analysis in [37] covers this wave-specific term. The numerical experiments always use contrast 10^3 and tau = 1e-3, so they do not test the contrast-dependence claim. The paper's headline claim therefore reduces to an unverified self-citation rather than to an in-paper derivation.

  2. self citation load bearing [Section 4, first paragraph and paragraph after (4.5)]
    "For detailed explanations and proofs, we refer the reader to [37]."

    The optimized decomposition used in all numerical examples is introduced in Section 4, and the section explicitly defers its proofs to the same overlapping-author preprint [37]. The subsequent claims 'It can be shown that gamma = 0' and the explicit stability bounds are stated without derivation. Since gamma appears in the factors (1 - gamma^2) of both stability conditions (3.26) and (3.29), the relaxed stability and decoupling used in the numerical results depend on a self-cited preprint, not on a proof contained in this paper.

full rationale

The paper's own stability analysis is not circular: Theorems 1 and 2 derive discrete energy conservation and the stability condition (3.26) directly from the scheme equations, and Theorem 3 states the analogous condition (3.29). These are genuine independent derivations. The circularity concern is concentrated in the contrast-independence claim, which is the advertised practical benefit. The sentence after Theorem 3 asserts that the m22-, a22-, and c22-norms are contrast-independent 'thanks to the construction of the cell problems [37]', and Section 4 likewise refers all proofs of the optimized decomposition to [37]. [37] is an overlapping-author arXiv preprint on flow problems, not machine-checked or otherwise externally verified in this paper. No in-paper computation or estimate shows the required contrast-independence, especially for the wave-specific c22 term that enters condition (3.29). Therefore the central claim that the stability time step is contrast-free rests on a load-bearing self-citation. This is not a self-definitional reduction: the stability inequalities themselves are derived in the paper, so the appropriate score is moderate rather than extreme.

Assumptions & free parameters 2 free parameters · 6 assumptions · 0 invented entities

The central derivation in this paper is the energy conservation and stability theorem for scheme 1, which is self-contained apart from the standard strengthened Cauchy-Schwarz inequality. Everything else that makes the method attractive is imported: the smoothness and truncation hypotheses of multicontinuum homogenization, the cancellation of mixed bilinear forms from [30], and most importantly the contrast-independence of the three norms, which is taken from the authors' own unpublished preprint [37]. No new physical entities are postulated; the continua and basis functions are inherited from prior multicontinuum theory.

free parameters (2)
  • Number of explicitly treated slow modes i0 = 1 in Examples 1 and 2, 2 in Examples 3 and 4
    The splitting dimension is selected by inspecting the generalized eigenvalues of (4.7) and picking the count of small eigenvalues by hand; no automatic criterion or sensitivity study is provided, and the stability estimates depend on lambda_{i0}.
  • Oversampling layer count l = l = ceil(-2 ln(H)), i.e., 5 for H = 1/10 and 6 for H = 1/20
    The oversampled region K+ is defined by extending K by l layers with this hand-chosen formula, carried over from references [8,18]; the paper does not test sensitivity to l.
assumptions (6)
  • domain assumption Macroscopic variables U_i are smooth over coarse regions
    Invoked before Eq. (2.6) to justify the two-term expansion and the dropping of terms of order H in (2.12).
  • domain assumption Two-term multicontinuum expansion u approximately phi_i U_i + phi_i^m grad_m U_i suffices
    Eq. (2.6) is the basis for the homogenized equation; the approximation error is not quantified in this paper.
  • domain assumption b_ij(U_j,W_i) + b_ji(W_i,U_j) = 0
    Used in Eq. (3.7) to remove mixed gradient terms; stated as verifiable by integration by parts and cited to [30].
  • ad hoc to paper For high/low continua, the m22, a22, and c22 norms are contrast-independent
    This is the load-bearing claim behind contrast-independent stability; no proof is given, only a citation to the authors' unpublished preprint [37].
  • ad hoc to paper Optimized decompositions are similar across coarse blocks, so the Rayleigh quotient can be localized
    Section 4, after Eq. (4.5): 'we assume that the decompositions for different coarse blocks are similar for simplicity'. No justification is provided.
  • standard math Strengthened Cauchy-Schwarz inequality holds with gamma in (0,1)
    Eq. (3.24) cites [39]; used in Theorems 2 and 3.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multicontinuum splitting schemes for multiscale wave problems." pith.science (2026). https://pith.science/paper/RR4RZRXC

@misc{pith2026250601670,
  author       = {Pith},
  title        = {Pith review of: Multicontinuum splitting schemes for multiscale wave problems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RR4RZRXC}},
  note         = {Machine review of arXiv:2506.01670}
}
read the original abstract

In this work, we propose multicontinuum splitting schemes for the wave equation with a high-contrast coefficient, extending our previous research on multiscale flow problems. The proposed approach consists of two main parts: decomposing the solution space into distinct components, and designing tailored time discretization schemes to enhance computational efficiency. To achieve the decomposition, we employ a multicontinuum homogenization method to introduce physically meaningful macroscopic variables and to separate fast and slow dynamics, effectively isolating contrast effects in high-contrast cases. This decomposition enables the design of schemes where the fast-dynamics (contrast-dependent) component is treated implicitly, while the slow-dynamics (contrast-independent) component is handled explicitly. The idea of discrete energy conservation is applied to derive the stability conditions, which are contrast-independent with appropriately chosen continua. We further discuss strategies for optimizing the space decomposition. These include a Rayleigh quotient problem involving tensors, and an alternative generalized eigenvalue decomposition to reduce computational effort. Finally, various numerical examples are presented to validate the accuracy and stability of our proposed method.

Figures

Figures reproduced from arXiv: 2506.01670 by the authors.

Figure 5.1
Figure 5.1. Left: Layered field κ in Example 1. Right: Reference solution at the final time T in Example 1 [PITH_FULL_IMAGE:figures/full_fig_p012_5_1.png] view at source ↗
Figure 5.2
Figure 5.2. Relative L 2 error for different schemes when H = 1/10 and l = 5 in Example 1. Left: e (1)(t). Right: e (2)(t) [PITH_FULL_IMAGE:figures/full_fig_p013_5_2.png] view at source ↗
Figure 5.3
Figure 5.3. Relative L 2 error for different schemes when H = 1/20 and l = 6 in Example 1. Left: e (1)(t). Right: e (2)(t). 13 [PITH_FULL_IMAGE:figures/full_fig_p013_5_3.png] view at source ↗
Figures from the paper (9 more)
Figure 5.4
Figure 5.4. Figure 5.4: Left: Point field κ in Example 2. Right: Reference solution at the final time T in Example 2 [PITH_FULL_IMAGE:figures/full_fig_p014_5_4.png]
Figure 5.5
Figure 5.5. Figure 5.5: Relative L 2 error for different schemes when H = 1/10 and l = 5 in Example 2. Left: e (1)(t). Right: e (2)(t) [PITH_FULL_IMAGE:figures/full_fig_p015_5_5.png]
Figure 5.6
Figure 5.6. Figure 5.6: Relative L 2 error for different schemes when H = 1/20 and l = 6 in Example 2. Left: e (1)(t). Right: e (2)(t). 15 [PITH_FULL_IMAGE:figures/full_fig_p015_5_6.png]
Figure 5.7
Figure 5.7. Figure 5.7: Left: Three-continuum field κ in Example 3. Right: Reference solution at the final time T in Example 3 [PITH_FULL_IMAGE:figures/full_fig_p016_5_7.png]
Figure 5.8
Figure 5.8. Figure 5.8: Relative L 2 error for different schemes when H = 1/10 and l = 5 in Example 3. Left: e (1)(t). Middle: e (2)(t). Right: e (3)(t) [PITH_FULL_IMAGE:figures/full_fig_p017_5_8.png]
Figure 5.9
Figure 5.9. Figure 5.9: Relative L 2 error for different schemes when H = 1/20 and l = 6 in Example 3. Left: e (1)(t). Middle: e (2)(t). Right: e (3)(t) [PITH_FULL_IMAGE:figures/full_fig_p017_5_9.png]
Figure 5.10
Figure 5.10. Figure 5.10: Left: Three-continuum field κ in Example 4. Right: Reference solution at the final time T in Example 4 [PITH_FULL_IMAGE:figures/full_fig_p017_5_10.png]
Figure 5.11
Figure 5.11. Figure 5.11: Relative L 2 error for different schemes when H = 1/10 and l = 5 in Example 4. Left: e (1)(t). Middle: e (2)(t). Right: e (3)(t) [PITH_FULL_IMAGE:figures/full_fig_p018_5_11.png]
Figure 5.12
Figure 5.12. Figure 5.12: Relative L 2 error for different schemes when H = 1/20 and l = 6 in Example 4. Left: e (1)(t). Middle: e (2)(t). Right: e (3)(t). 6 Conclusions In this work, we propose multicontinuum splitting schemes for the wave equation with a high-contrast coefficient. This is …

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

39 extracted references · 36 canonical work pages

  1. [37]

    Multicontinuum splitting scheme for multiscale flow problems

    Yalchin Efendiev, Wing Tat Leung, Buzheng Shan, and Min Wang. Multicontinuum splitting scheme for multiscale flow problems. arXiv preprint arXiv:2410.05253 , 2024

    arXiv 2024

  2. [1]

    Sh-wave propagation in heterogeneous media: Velocity-stress finite-difference method

    Jean Virieux. Sh-wave propagation in heterogeneous media: Velocity-stress finite-difference method. Geophysics, 49(11):1933–1942, 1984. 18

    work page 1933

  3. [2]

    P-sv wave propagation in heterogeneous media: Velocity-stress finite-difference method

    Jean Virieux. P-sv wave propagation in heterogeneous media: Velocity-stress finite-difference method. Geophysics, 51(4):889–901, 1986

    work page 1986

  4. [3]

    Abdulle and P

    A. Abdulle and P. Henning. Multiscale methods for wave problems in heterogeneous media. In R´ emi Abgrall and Chi-Wang Shu, editors,Handbook of Numerical Methods for Hyperbolic Problems, volume 18 of Handbook of Numerical Analysis , pages 545–576. Elsevier, 2017

    work page 2017

  5. [4]

    Asymptotic analysis for periodic structures

    George Papanicolau, Alain Bensoussan, and J-L Lions. Asymptotic analysis for periodic structures . Elsevier, 1978

    work page 1978

  6. [5]

    Analysis of upscaling absolute permeability

    Xiao-Hui Wu, Yalchin Efendiev, and Thomas Y Hou. Analysis of upscaling absolute permeability. Discrete and Continuous Dynamical Systems Series B , 2(2):185–204, 2002

    work page 2002

  7. [6]

    Numerical homogenization of the acoustic wave equations with a continuum of scales

    Houman Owhadi and Lei Zhang. Numerical homogenization of the acoustic wave equations with a continuum of scales. Computer Methods in Applied Mechanics and Engineering , 198(3-4):397–406, 2008

    work page 2008

  8. [7]

    Homogenization and porous media , volume 6

    Ulrich Hornung. Homogenization and porous media , volume 6. Springer Science & Business Media, 2012

    work page 2012

Show all 39 references
  1. [8]

    A multiscale finite element method for elliptic problems in composite materials and porous media

    Thomas Y Hou and Xiao-Hui 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

  2. [9]

    Convergence of a multiscale finite element method for elliptic problems with rapidly oscillating coefficients

    Thomas Hou, Xiao-Hui Wu, and Zhiqiang Cai. Convergence of a multiscale finite element method for elliptic problems with rapidly oscillating coefficients. Mathematics of computation, 68(227):913–943, 1999

    1999

  3. [10]

    Multiscale finite element methods: theory and applications , volume 4

    Yalchin Efendiev and Thomas Y Hou. Multiscale finite element methods: theory and applications , volume 4. Springer Science & Business Media, 2009

    2009

  4. [11]

    Analysis of global multiscale finite element meth- ods for wave equations with continuum spatial scales.Applied numerical mathematics, 60(8):862–876, 2010

    Lijian Jiang, Yalchin Efendiev, and Victor Ginting. Analysis of global multiscale finite element meth- ods for wave equations with continuum spatial scales.Applied numerical mathematics, 60(8):862–876, 2010

    2010

  5. [12]

    Localization of elliptic multiscale problems

    Axel M ˚ alqvist and Daniel Peterseim. Localization of elliptic multiscale problems. Mathematics of Computation, 83(290):2583–2603, 2014

    2014

  6. [13]

    Localized orthogonal decomposition techniques for boundary value problems

    Patrick Henning and Axel M ˚ alqvist. Localized orthogonal decomposition techniques for boundary value problems. SIAM Journal on Scientific Computing , 36(4):A1609–A1634, 2014

    2014

  7. [14]

    Localized orthogonal decomposition method for the wave equation with a continuum of scales

    Assyr Abdulle and Patrick Henning. Localized orthogonal decomposition method for the wave equation with a continuum of scales. Mathematics of Computation , 86(304):549–587, 2017

    2017

  8. [15]

    Generalized multiscale finite element methods (gmsfem)

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

    2013

  9. [16]

    Generalized multiscale finite element methods for wave propagation in heterogeneous media

    Eric T Chung, Yalchin Efendiev, and Wing Tat Leung. Generalized multiscale finite element methods for wave propagation in heterogeneous media. Multiscale Modeling & Simulation , 12(4):1691–1721, 2014

    2014

  10. [17]

    Generalized multiscale finite-element method (gmsfem) for elastic wave propagation in heterogeneous, anisotropic media

    Kai Gao, Shubin Fu, Richard L Gibson Jr, Eric T Chung, and Yalchin Efendiev. Generalized multiscale finite-element method (gmsfem) for elastic wave propagation in heterogeneous, anisotropic media. Journal of Computational Physics , 295:161–188, 2015

    2015

  11. [18]

    Chung, Yalchin Efendiev, and Wing Tat Leung

    Eric T. Chung, Yalchin Efendiev, and Wing Tat Leung. Constraint energy minimizing general- ized multiscale finite element method. Computer Methods in Applied Mechanics and Engineering , 339:298–319, 2018. 19

    2018

  12. [19]

    Explicit and energy- conserving constraint energy minimizing generalized multiscale discontinuous galerkin method for wave propagation in heterogeneous media

    Siu Wun Cheung, Eric T Chung, Yalchin Efendiev, and Wing Tat Leung. Explicit and energy- conserving constraint energy minimizing generalized multiscale discontinuous galerkin method for wave propagation in heterogeneous media. Multiscale Modeling & Simulation, 19(4):1736–1759, 2021

    2021

  13. [20]

    Multiscale Model Reduction: Multiscale Finite Element Methods and Their Generalizations , volume 212

    Eric Chung, Yalchin Efendiev, and Thomas Y Hou. Multiscale Model Reduction: Multiscale Finite Element Methods and Their Generalizations , volume 212. Springer Nature, 2023

    2023

  14. [21]

    Heterogeneous multiscale methods: a review

    E Weinan, Bjorn Engquist, Xiantao Li, Weiqing Ren, and Eric Vanden-Eijnden. Heterogeneous multiscale methods: a review. Communications in computational physics , 2(3):367–450, 2007

    2007

  15. [22]

    The heterogeneous multiscale method

    Assyr Abdulle, E Weinan, Bj¨ orn Engquist, and Eric Vanden-Eijnden. The heterogeneous multiscale method. Acta Numerica, 21:1–87, 2012

    2012

  16. [23]

    Finite element heterogeneous multiscale method for the wave equation

    Assyr Abdulle and Marcus J Grote. Finite element heterogeneous multiscale method for the wave equation. Multiscale Modeling & Simulation , 9(2):766–792, 2011

    2011

  17. [24]

    The variational mul- tiscale method—a paradigm for computational mechanics

    Thomas JR Hughes, Gonzalo R Feij´ oo, Luca Mazzei, and Jean-Baptiste Quincy. The variational mul- tiscale method—a paradigm for computational mechanics. Computer methods in applied mechanics and engineering, 166(1-2):3–24, 1998

    1998

  18. [25]

    Chung, Yalchin Efendiev, Wing Tat Leung, Maria Vasilyeva, and Yating Wang

    Eric T. Chung, Yalchin Efendiev, Wing Tat Leung, Maria Vasilyeva, and Yating Wang. Non-local multi-continua upscaling for flows in heterogeneous fractured media. Journal of Computational Physics, 372:22–34, 2018

    2018

  19. [26]

    Nonlocal multicontinua upscaling for multicontinua flow problems in fractured porous media

    Maria Vasilyeva, Eric T Chung, Siu Wun Cheung, Yating Wang, and Georgy Prokopev. Nonlocal multicontinua upscaling for multicontinua flow problems in fractured porous media. Journal of Computational and Applied Mathematics , 355:258–267, 2019

    2019

  20. [27]

    Deep multiscale model learning

    Yating Wang, Siu Wun Cheung, Eric T Chung, Yalchin Efendiev, and Min Wang. Deep multiscale model learning. Journal of Computational Physics , 406:109071, 2020

    2020

  21. [28]

    Hybrid explicit–implicit learning for multiscale problems with time dependent source

    Yalchin Efendiev, Wing Tat Leung, Wenyuan Li, and Zecheng Zhang. Hybrid explicit–implicit learning for multiscale problems with time dependent source. Communications in Nonlinear Science and Numerical Simulation , 120:107081, 2023

    2023

  22. [29]

    Locally subspace-informed neural operators for efficient mul- tiscale pde solving

    Alexander Rudikov, Vladimir Fanaskov, Sergei Stepanov, Buzheng Shan, Ekaterina Muravleva, Yalchin Efendiev, and Ivan Oseledets. Locally subspace-informed neural operators for efficient mul- tiscale pde solving. arXiv preprint arXiv:2505.16030 , 2025

    2025

  23. [30]

    Multicontinuum homogenization and its relation to nonlocal multicontinuum theories

    Yalchin Efendiev and Wing Tat Leung. Multicontinuum homogenization and its relation to nonlocal multicontinuum theories. Journal of Computational Physics , 474:111761, 2023

    2023

  24. [31]

    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

    2024

  25. [32]

    Multicontinuum homogeniza- tion for coupled flow and transport equations

    Dmitry Ammosov, Jian Huang, Wing Tat Leung, and Buzheng Shan. Multicontinuum homogeniza- tion for coupled flow and transport equations. Journal of Computational and Applied Mathematics , 471:116736, 2026

    2026

  26. [33]

    Multicontinuum homogenization in perforated domains

    Wei Xie, Yalchin Efendiev, Yunqing Huang, Wing Tat Leung, and Yin Yang. Multicontinuum homogenization in perforated domains. Journal of Computational Physics , 530:113845, 2025

    2025

  27. [34]

    Contrast-independent partially explicit time discretizations for multiscale flow problems.Journal of Computational Physics, 445:110578, 2021

    Eric T Chung, Yalchin Efendiev, Wing Tat Leung, and Petr N Vabishchevich. Contrast-independent partially explicit time discretizations for multiscale flow problems.Journal of Computational Physics, 445:110578, 2021. 20

    2021

  28. [35]

    Contrast-independent partially explicit time discretizations for multiscale wave problems

    Eric T Chung, Yalchin Efendiev, Wing Tat Leung, and Petr N Vabishchevich. Contrast-independent partially explicit time discretizations for multiscale wave problems. Journal of Computational Physics, 466:111226, 2022

    2022

  29. [36]

    Contrast-independent, partially- explicit time discretizations for nonlinear multiscale problems

    Eric T Chung, Yalchin Efendiev, Wing Tat Leung, and Wenyuan Li. Contrast-independent, partially- explicit time discretizations for nonlinear multiscale problems. Mathematics, 9(23):3000, 2021

    2021

  30. [38]

    Some convergence analysis for multicontinuum homogenization

    Wing Tat Leung. Some convergence analysis for multicontinuum homogenization. arXiv preprint arXiv:2401.12799, 2024

    2024 arXiv

  31. [39]

    Strengthened cauchy-schwarz and h¨ older inequalities

    JM Aldaz. Strengthened cauchy-schwarz and h¨ older inequalities. JIPAM. Journal of Inequalities in Pure & Applied Mathematics [electronic only] , 10(4):Paper–No, 2009. 21

    2009

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.