Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Numerical approximation of effective diffusivities in homogenization of nondivergence-form equations with large drift by a Lagrangian method

T0 review · 3 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 establishes a Lagrangian Monte Carlo method that computes effective diffusivities in nondivergence-form homogenization with large drift to second order in the time step, with total error O(h^2 + 1/T + 1/sqrt(M)).

desk verdict Useful Lagrangian method for effective diffusivity with a repairable but real gap in the main T-uniform error estimate. read the letter →

arxiv 2506.14073 v1 pith:SWOF6FBG submitted 2025-06-17 math.NA cs.NA

Timo Sprekeler , Han Wu , Zhiwen Zhang This is my paper · ORCID
classification math.NAcs.NA MSC 35B2760H3565M1265M75
keywords numericalhomogenizationnondivergence-formequationsmodifiedequationapproachMonteCarlomethodeffectivediffusivityconvergenceanalysislargedriftLagrangian
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's central claim is that the effective diffusion matrix of a periodic nondivergence-form elliptic equation with large drift can be computed by directly simulating the associated diffusion process and taking the long-time variance of particle paths, without solving the Fokker-Planck-Kolmogorov equation. It proves a Lagrangian-Eulerian equivalence that holds even when the drift is not centered with respect to the invariant measure, which is exactly the situation produced by numerical discretizations. The proposed modified Milstein scheme, built through modified equations, converges weakly at second order and yields an empirical variance estimator with error bounded by C($h^{2}$ + 1/T + 1/$\sqrt$(M)) uniformly in T and M. If this bound is right, the method offers a high-dimensional, sampling-based route to effective diffusivities that improves on the first-order Euler-Maruyama scheme, which is known to fail in this regime.

What carries the argument

The central object is the modified Milstein integrator (3.5), whose drift and diffusion coefficients are b_h = b + h b_1 and sigma_h = $\sigma$ + h sigma_1 with the first-order corrections (3.6) obtained from the modified equation approach. The argument is carried by backward error analysis: the discrete scheme is replaced by a modified SDE whose generator, invariant measure r'_h, and cell-problem corrector chi'_h are all O($h^{2}$) close to those of the original SDE, through Lemma 4.1 and Lemma 4.2. Theorem 2.1 supplies the load-bearing identity that the normalized variance tends to the Eulerian effective diffusivity even without centering, so the modified flow's variance is the right quantity to estimate.

What would settle it

Take the modified Milstein scheme at a fixed small h, run it for a sequence of increasing final times T (say T = 10, 100, 1000) with large M, and measure whether the $h^{2}$ part of the error in Var(x_T)/(2T) stays bounded as T grows; a growing $h^{2}$ contribution would contradict the uniformity claimed in Corollary 4.2. A complementary check is to verify the backward error estimate (4.6) for the moment functions x and x x^T at times t_n up to large T with numerically estimated constants.

Watch

Extended reading notes

Core claim

The discovery is that the long-time variance of the rescaled diffusion X_t, normalized as Var(X_T)/(2T), converges to the same effective diffusion matrix A^L as the Eulerian cell-problem formula, even when the centering condition \bar b = 0 fails, and that a Milstein scheme with modified drift and diffusion coefficients approximates this limit with second-order weak error. The main quantitative statement is Corollary 4.2: for the modified Milstein scheme (3.5), the empirical variance estimator satisfies |Var(x_{t_N})/(2T) - A^L| <= C($h^{2}$ + 1/T + 1/$\sqrt$(M)) for small h, with the constant independent of h, T, and M. The proof combines the weak convergence of the invariant measure (Corollary 4.1), an O($h^{2}$) closeness of the backward-error modified SDE (estimate (4.6)), and an O($h^{2}$) bound on the modified effective diffusivity A^L_h (Theorem 4.1).

Load-bearing premise

The load-bearing premise is that the constants in the finite-time weak error estimate (Proposition 4.1) and the backward error estimate (4.5) are independent of the final time T; Corollary 4.2 divides the finite-time error by T, and if those constants grow with T the stated O($h^{2}$ + 1/T + 1/$\sqrt$(M)) bound would no longer follow.

Editorial extensions

If this is right

  • Effective diffusivities in nondivergence-form homogenization with large drift can be computed by direct SDE simulation, bypassing the cost of solving the Fokker-Planck-Kolmogorov equation.
  • The total error of the empirical variance estimator is O(h^2 + 1/T + 1/sqrt(M)), so the time step, simulation horizon, and particle number can be balanced to reach a target accuracy.
  • The Lagrangian-Eulerian equivalence holds without the centering condition, which explains why standard Euler-Maruyama can fail and why a scheme that preserves the diffusion structure is needed.
  • Numerical experiments in two and three dimensions show second-order convergence for the modified Milstein scheme, whereas Euler-Maruyama stays at first order.

Reading between the lines

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

  • If the T-uniformity of the constants holds, the error bound suggests tuning h ~ T^{-1/2} and M ~ T^2 to equalize the three error contributions, a rule the paper does not state explicitly.
  • The same modified-equation construction could be applied to other long-time observables of the diffusion, such as occupation measures or mean residence times, wherever the invariant measure is the target.
  • A direct testable consequence is that at fixed small T the h^2 term enters as C h^2/T, so numerically the observed convergence in h may look faster than second order until T is large.
  • The framework is a candidate for time-dependent or random periodic media, where an Eulerian Fokker-Planck solver is much harder to build but particle simulation remains straightforward.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes a Lagrangian (particle/Monte Carlo) method for computing the effective diffusivity in periodic homogenization of nondivergence-form elliptic equations with large drift. The authors observe that the effective diffusivity can be characterized by the long-time variance of a rescaled diffusion process, simulate that process with a modified Milstein scheme built from the modified-equation framework of Abdulle et al., and estimate the variance from an ensemble of particles. The paper proves a Lagrangian--Eulerian equivalence (Theorem 2.1), gives a modified-equation construction for a second-order weak scheme (Section 3), derives an O(h^2) error for the modified effective diffusivity (Theorem 4.1), and claims an O(h^2 + 1/T + 1/sqrt(M)) bound for the empirical variance estimator (Corollary 4.2). Numerical experiments in 2D and 3D show second-order convergence in the time step and support the practical value of the scheme.

Significance. If the main error bound can be made rigorous, the method is significant: it provides a Monte Carlo route to effective diffusivities in nondivergence-form homogenization that avoids solving the Fokker--Planck--Kolmogorov equation, is dimension-friendly, and offers second-order weak convergence through a modified Milstein scheme. The paper also gives a useful explanation for the poor performance of standard Euler--Maruyama schemes reported in earlier work, attributing it to the failure of the centering condition for modified equations. The numerical experiments are extensive and convincingly demonstrate second-order convergence of the proposed scheme for constant and non-constant diffusion coefficients in 2D and 3D. The main weaknesses are analytic: two load-bearing steps in the proofs are not justified as stated, and the uniform-in-time error estimates needed for Corollary 4.2 are not established.

major comments (3)
  1. [§4.2, Corollary 4.2] The proof of (4.15) applies the weak error estimate (4.6) at t_N = T to the moment functions f_1(x)=x and f_2(x)=xx^T, treating the constants K_1, K_2 as independent of T. However, Proposition 4.1 and the backward error estimate (4.5)--(4.6) are stated only for a fixed time t_n, with no proof that the constants are uniform in t_n. For unbounded test functions and a non-centered drift, the natural scaling is |E[\tilde X_T]-E[X'_T]| = O(h^2 T) and |E[\tilde X_T \tilde X_T^T]-E[X'_T X'_T^T]| = O(h^2 T^2) before variance cancellation, which would turn the claimed O(h^2) contribution into O(h^2 T) after division by 2T. The proof does not exploit the cancellation between E|X|^2 and |E X|^2 in Var(X)=E|X|^2-|E X|^2. This T-uniformity is the load-bearing premise for the main O(h^2 + 1/T + 1/sqrt(M)) claim; without it, the estimate (4.15) is unsupported. The authors should either prove T-uniform weak error bounds for the relevant moments, prove a variance-specific cancellation estimate, or weaken the stated bound accordingly.
  2. [§2, Eq. (2.6)] The final step of the proof of Theorem 2.1 asserts that the product (∫_0^t (I+∇χ'(X_s))σ(X_s)dW_s)(χ'(X_t)-E[χ'(X_t)])^T is a martingale whose expectation vanishes. This is not generally true: the Itô integral is a martingale, but multiplying by a mean-zero non-martingale does not preserve the martingale property. For example, with H≡1 and Y_t=W_t, E[∫_0^t dW_s · W_t]=E[W_t^2]=t≠0. The estimate (2.6) therefore rests on an invalid assertion. The result is likely repairable by applying Cauchy--Schwarz, which gives a cross-term of order O(1/sqrt{t}) after division by t, but the proof as written is incorrect.
  3. [§4.2, first term of (4.16)] The proof of Corollary 4.2 bounds the Monte Carlo sampling error by L_1/sqrt{M} with L_1 independent of T. This is asserted from the central limit theorem, but the variance of the sample variance at time T generally depends on T, and no argument is given for its uniformity. A similar gap appears in Theorem 4.1, where the uniform L^4 bound on ∥∇χ'_h∥_{L^4(T^d)} is asserted by citing W^{2,p} estimates without proof. Both points need to be made rigorous for the stated uniform-in-T/M estimate (4.15).
minor comments (5)
  1. [§3.2, after Eq. (3.6)] The displayed formulas for b_1 and σ_1 end with a comma instead of a period, and the formulas for b'_1 and σ'_1 in §4.2 contain terms proportional to h inside the h^2 correction; this appears dimensionally inconsistent and should be checked against [1].
  2. [§2, proof of Theorem 2.1] The notation 'r_s denotes the probability density of X_s' is imprecise: since X_s is an R^d-valued process with periodic coefficients, the integrals over T^d require r_s to be the density of X_s modulo Z^d (the folded density), not the density on R^d. This should be stated explicitly.
  3. [§4.2, Corollary 4.2] The notation 'let x_{t_N}=(x_{t_N,[1]},...,x_{t_N,[d]})^T be M independent realizations, where each x_{t_N,[i]}∈R^M' is confusing; presumably x_{t_N,[j]} is the vector of M samples of the j-th component, and this should be rewritten for clarity.
  4. [§3.2, non-commutative case] The statement 'Without altering the weak convergence rate, we can simply take q=2' needs a reference or a brief justification, since the truncation error of the Fourier--Legendre series enters the weak error analysis and could, in principle, affect the claimed second-order rate.
  5. [§5.1, Figure 1] Figure 1 reports second-order convergence for a modified Euler--Maruyama scheme. This is consistent with Remark 4.1 only because the diffusion coefficient is constant in that example; the text should state this explicitly to avoid an apparent contradiction.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the modified-Milstein coefficients are taken from external modified-equation theory, and the error analysis compares independent variance terms without fitting A^L.

full rationale

The derivation chain is self-contained in the relevant sense. The paper's target quantity A^L is characterized by the long-time variance of the diffusion (Theorem 2.1), and the numerical estimator is the empirical variance of a modified Milstein process. The modified drift and diffusion corrections b_h = b + h b_1 and sigma_h = sigma + h sigma_1 in (3.6) are taken from the external modified-equation theory of Abdulle, Cohen, Vilmart, and Zygalakis [1], not fitted to A^L or to any effective-diffusivity data; their weak second-order property is quoted as a special case of [1, Theorem 2.1]. The convergence proof in Corollary 4.2 is a triangle-inequality comparison of four separate errors: Monte Carlo sampling, modified-Milstein versus modified-SDE moments, modified-SDE variance versus A^L_h, and A^L_h versus A^L. None of these steps defines the target in terms of the estimator or fits a parameter to the output. The self-citations (e.g., [11], [15], [40], [41]) are background on finite-element homogenization and are not load-bearing for the new Lagrangian estimator. The main weaknesses are correctness risks rather than circularity: Corollary 4.2 applies the weak-error estimate (4.6) at t_N = T and divides by T without proving that the constants K_1 and K_2 are independent of T, and the proof of Theorem 2.1 contains an unsubstantiated martingale assertion for the cross term in (2.6). These affect the validity of the stated O(h^2 + 1/T + 1/sqrt(M)) bound, but they do not make the derivation circular. Accordingly, the circularity score is 0.

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

The method introduces no fitted constants. h, T and M are discretization parameters; q=2 is a fixed truncation choice. The analysis borrows the modified equation and backward error analysis machinery from prior literature, which is the main external axiom dependence.

assumptions (6)
  • domain assumption Uniform ellipticity and smoothness of A and b; invariant density r is positive, smooth and exponentially mixing as in (1.6).
    Used in Theorem 2.1 and throughout to justify invariant measure, cell problems and the variance limits.
  • standard math Fredholm alternative and elliptic regularity for the cell problem Lchi = b - <b>_r on the torus.
    Invoked in (1.7), (2.2), Lemma 4.1 and Lemma 4.2 to guarantee existence, uniqueness and stability.
  • domain assumption Modified equation framework of Abdulle, Cohen, Vilmart and Zygalakis [1], in particular Proposition 4.1 and the form of the leading error operator L1.
    Section 3.2 and Section 4.1 rely on [1] for the construction of b1 and sigma1 and for the second-order weak convergence claim.
  • domain assumption Backward error analysis estimates (4.5) from [12, 39] for Milstein-type integrators.
    Used in Section 4.2 to define the second-order modified continuous-time process X'_t and to justify the h^2 error between the modified integrator and X'_t.
  • standard math W^{2,p} estimates for the cell problem (4.9) to control ||grad chi'_h||_{L4} uniformly in h.
    Needed in Theorem 4.1 to close the error bound for the effective diffusivity.
  • ad hoc to paper Assumption 1: the numerical solution admits a weak Taylor expansion U_{b,sigma}(phi,x,h) = phi + hLphi + sum h^{j+1} A_j phi.
    Stated in Section 4.1 and used in Proposition 4.1; it is not proved for the Milstein scheme in this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Numerical approximation of effective diffusivities in homogenization of nondivergence-form equations with large drift by a Lagrangian method." pith.science (2026). https://pith.science/paper/SWOF6FBG

@misc{pith2026250614073,
  author       = {Pith},
  title        = {Pith review of: Numerical approximation of effective diffusivities in homogenization of nondivergence-form equations with large drift by a Lagrangian method},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SWOF6FBG}},
  note         = {Machine review of arXiv:2506.14073}
}
read the original abstract

In this paper, we study numerical methods for the homogenization of linear second-order elliptic equations in nondivergence-form with periodic diffusion coefficients and large drift terms. Upon noting that the effective diffusion matrix can be characterized through the long-time variance of an associated diffusion process, we construct a Lagrangian numerical scheme based on a direct simulation of the underlying stochastic differential equation and utilizing the framework of modified equations, thereby avoiding the need to solve the Fokker--Planck--Kolmogorov equation. Through modified equation analysis, we derive higher-order weak convergence rates for our method. Finally, we conduct numerical experiments to demonstrate the accuracy of the proposed method. The results show that the method efficiently computes effective diffusivities, even in high dimensions.

Figures

Figures reproduced from arXiv: 2506.14073 by the authors.

Figure 1
Figure 1. Convergence rate of [PITH_FULL_IMAGE:figures/full_fig_p017_1.png] view at source ↗
Figure 2
Figure 2. Convergence rate of Frobenius norm via Euler–Maruyama and modified Milstein [PITH_FULL_IMAGE:figures/full_fig_p018_2.png] view at source ↗
Figure 3
Figure 3. Convergence rate of A11 of Euler–Maruyama and modified Milstein [PITH_FULL_IMAGE:figures/full_fig_p019_3.png] view at source ↗
Figures from the paper (3 more)
Figure 5
Figure 5. Figure 5: Convergence rate of Frobenius norm via Euler–Maruyama and modified Milstein 6. Conclusion In this paper, we have developed a probabilistic numerical framework for accurately computing the effective diffusion matrix in periodic homogenization of nondivergence-form ellip…
Figure 6
Figure 6. Figure 6: Convergence rate of A11 of Euler–Maruyama and modified Milstein [PITH_FULL_IMAGE:figures/full_fig_p021_6.png]
Figure 8
Figure 8. Figure 8: heat map of exact density and numerical density for the invariant measure efficient, accurate approach for homogenization problems with large drift, with potential applications to turbulent transport in fluid dynamics and anomalous diffusion in biological systems. Ackn…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. A Unified Kullback--Leibler Divergence Analysis of Generative Diffusion Models via Entropy Production Rate

    math.NA 2026-08 conditional novelty 5.0 of 10

    Diffusion model generation error is decomposed through an entropy-production-rate identity that claims O(h²) Euler–Maruyama KL bounds and unifies score SDE, PF-ODE, flow matching, and stochastic interpolant analyses.

Reference graph

Works this paper leans on

46 extracted references · 46 canonical work pages · cited by 1 Pith paper

  1. [1]

    Abdulle, D

    A. Abdulle, D. Cohen, G. Vilmart, and K. C. Zygalakis. High weak order methods for stochastic differential equations based on modified equations.SIAM J. Sci. Comput., 34(3):A1800–A1823, 2012

    work page 2012

  2. [2]

    Abdulle, G

    A. Abdulle, G. Vilmart, and K. C. Zygalakis. High order numerical approximation of the invariant measure of ergodic SDEs.SIAM J. Numer. Anal., 52(4):1600–1622, 2014

    work page 2014

  3. [3]

    Armstrong, B

    S. Armstrong, B. Fehrman, and J. Lin. Green function and invariant measure estimates for nondivergence form elliptic homogenization, 2022

    work page 2022

  4. [4]

    Armstrong and J

    S. Armstrong and J. Lin. Optimal quantitative estimates in stochastic homogenization for elliptic equations in nondivergence form.Arch. Ration. Mech. Anal., 225(2):937–991, 2017

    work page 2017

  5. [5]

    S. N. Armstrong and C. K. Smart. Quantitative stochastic homogenization of elliptic equations in nondivergence form.Arch. Ration. Mech. Anal., 214(3):867–911, 2014

    work page 2014

  6. [6]

    Bensoussan, J.-L

    A. Bensoussan, J.-L. Lions, and G. Papanicolaou.Asymptotic analysis for periodic structures. AMS Chelsea Publishing, Providence, RI, 2011. Corrected reprint of the 1978 original

    work page 2011

  7. [7]

    R. A. Biezemans, C. Le Bris, F. Legoll, and A. Lozinski. MsFEM for advection-dominated problems in hetero- geneous media: Stabilization via nonconforming variants.Computer Methods in Applied Mechanics and Engi- neering, 433, Part A, 2025. Article 117496

    work page 2025

  8. [8]

    V. I. Bogachev and S. V. Shaposhnikov. Integrability and continuity of solutions to double divergence form equations.Ann. Mat. Pura Appl. (4), 196(5):1609–1635, 2017

    work page 2017

Show all 46 references
  1. [9]

    Bonizzoni, P

    F. Bonizzoni, P. Freese, and D. Peterseim. Super-localized orthogonal decomposition for convection-dominated diffusion problems.BIT, 64(3):Paper No. 33, 2024

    2024

  2. [10]

    Capdeboscq

    Y. Capdeboscq. Homogenization of a diffusion equation with drift.C. R. Acad. Sci. Paris S´ er. I Math., 327(9):807–812, 1998

    1998

  3. [11]

    Capdeboscq, T

    Y. Capdeboscq, T. Sprekeler, and E. S¨ uli. Finite element approximation of elliptic homogenization problems in nondivergence-form.ESAIM Math. Model. Numer. Anal., 54(4):1221–1257, 2020

    2020

  4. [12]

    Debussche and E

    A. Debussche and E. Faou. Weak backward error analysis for SDEs.SIAM J. Numer. Anal., 50(3):1735–1752, 2012

    2012

  5. [13]

    Engquist and P

    B. Engquist and P. E. Souganidis. Asymptotic and numerical homogenization.Acta Numer., 17:147–190, 2008

    2008

  6. [14]

    Fannjiang and G

    A. Fannjiang and G. Papanicolaou. Convection enhanced diffusion for periodic flows.SIAM J. Appl. Math., 54(2):333–408, 1994

    1994

  7. [15]

    Freese, D

    P. Freese, D. Gallistl, D. Peterseim, and T. Sprekeler. Computational multiscale methods for nondivergence-form elliptic partial differential equations.Comput. Methods Appl. Math., 24(3):649–672, 2024

    2024

  8. [16]

    Gallistl, T

    D. Gallistl, T. Sprekeler, and E. S¨ uli. Mixed Finite Element Approximation of Periodic Hamilton–Jacobi– Bellman Problems With Application to Numerical Homogenization.Multiscale Model. Simul., 19(2):1041–1065, 2021

    2021

  9. [17]

    J. Garnier. Homogenization in a periodic and time-dependent potential.SIAM J. Appl. Math., 57(1):95–111, 1997

    1997

  10. [18]

    X. Guo, T. Sprekeler, and H. V. Tran. Characterizations of diffusion matrices in homogenization of elliptic equations in nondivergence-form.Calc. Var. Partial Differential Equations, 64(1), 2025. Paper No. 1

    2025

  11. [19]

    Guo and H

    X. Guo and H. V. Tran. Optimal convergence rates in stochastic homogenization in a balanced random environ- ment, 2023

    2023

  12. [20]

    Guo and H

    X. Guo and H. V. Tran. Stochastic integrability of heat-kernel bounds for random walks in a balanced random environment.Electron. J. Probab., 29, 2024. Paper No. 194

    2024

  13. [21]

    X. Guo, H. V. Tran, and Y. Yu. Remarks on optimal rates of convergence in periodic homogenization of linear elliptic equations in non-divergence form.Partial Differ. Equ. Appl., 1(4), 2020. Paper No. 15

    2020

  14. [22]

    Hairer and E

    M. Hairer and E. Pardoux. Homogenization of periodic linear degenerate PDEs.J. Funct. Anal., 255(9):2462– 2487, 2008

    2008

  15. [23]

    Henning and M

    P. Henning and M. Ohlberger. The heterogeneous multiscale finite element method for advection-diffusion prob- lems with rapidly oscillating coefficients and large expected drift.Netw. Heterog. Media, 5(4):711–744, 2010. NUMERICAL APPROXIMATION OF EFFECTIVE DIFFUSIVITIES 23

    2010

  16. [24]

    Henning and M

    P. Henning and M. Ohlberger. A note on homogenization of advection-diffusion problems with large expected drift.Z. Anal. Anwend., 30(3):319–339, 2011

    2011

  17. [25]

    V. V. Jikov, S. M. Kozlov, and O. A. Ole ˘ ınik.Homogenization of differential operators and integral functionals. Springer-Verlag, Berlin, 1994. Translated from the Russian

    1994

  18. [26]

    Jing and Y

    W. Jing and Y. Zhang. On the periodic homogenization of elliptic equations in nondivergence form with large drifts.Multiscale Model. Simul., 21(4):1486–1501, 2023

    2023

  19. [27]

    E. L. Kawecki and T. Sprekeler. Discontinuous Galerkin andC 0-IP finite element approximation of periodic Hamilton-Jacobi-Bellman-Isaacs problems with application to numerical homogenization.ESAIM Math. Model. Numer. Anal., 56(2):679–704, 2022

    2022

  20. [28]

    Kim and K.-A

    S. Kim and K.-A. Lee. Higher order convergence rates in theory of homogenization: equations of non-divergence form.Arch. Ration. Mech. Anal., 219(3):1273–1304, 2016

    2016

  21. [29]

    D. F. Kuznetsov. Application of multiple Fourier-Legendre series to implementation of strong exponential Mil- stein and Wagner-Platen methods for non-commutative semilinear stochastic partial differential equations.Dif- fer. Uravn. Protsessy Upr., (3):129–162, 2020

    2020

  22. [30]

    Landim, S

    C. Landim, S. Olla, and H. T. Yau. Convection-diffusion equation with space-time ergodic random flow.Probab. Theory Related Fields, 112(2):203–220, 1998

    1998

  23. [31]

    Le Bris, F

    C. Le Bris, F. Legoll, and F. Madiot. A numerical comparison of some multiscale finite element approaches for advection-dominated problems in heterogeneous media.ESAIM Math. Model. Numer. Anal., 51(3):851–888, 2017

    2017

  24. [32]

    G. Li, D. Peterseim, and M. Schedensack. Error analysis of a variational multiscale stabilization for convection- dominated diffusion equations in two dimensions.IMA J. Numer. Anal., 38(3):1229–1253, 2018

    2018

  25. [33]

    G. N. Milstein and M. V. Tretyakov.Stochastic numerics for mathematical physics. Scientific Computation. Springer, Cham, second edition, 2021

    2021

  26. [34]

    G. A. Pavliotis and A. M. Stuart. Periodic homogenization for inertial particles.Phys. D, 204(3-4):161–187, 2005

    2005

  27. [35]

    G. A. Pavliotis and A. M. Stuart.Multiscale methods, volume 53 ofTexts in Applied Mathematics. Springer, New York, 2008. Averaging and homogenization

    2008

  28. [36]

    G. A. Pavliotis, A. M. Stuart, and L. Band. Monte Carlo studies of effective diffusivities for inertial particles. InMonte Carlo and quasi-Monte Carlo methods 2004, pages 431–441. Springer, Berlin, 2006

    2004

  29. [37]

    G. A. Pavliotis, A. M. Stuart, and K. C. Zygalakis. Calculating effective diffusiveness in the limit of vanishing molecular diffusion.J. Comput. Phys., 228(4):1030–1055, 2009

    2009

  30. [38]

    J. Qian, T. Sprekeler, H. V. Tran, and Y. Yu. Optimal Rate of Convergence in Periodic Homogenization of Viscous Hamilton-Jacobi Equations.Multiscale Model. Simul., 22(4):1558–1584, 2024

    2024

  31. [39]

    S. Reich. Backward error analysis for numerical integrators.SIAM J. Numer. Anal., 36(5):1549–1570, 1999

    1999

  32. [40]

    Sprekeler

    T. Sprekeler. Homogenization of nondivergence-form elliptic equations with discontinuous coefficients and finite element approximation of the homogenized problem.SIAM J. Numer. Anal., 62(2):646–666, 2024

    2024

  33. [41]

    Sprekeler, E

    T. Sprekeler, E. S¨ uli, and Z. Zhang. Finite element approximation of stationary Fokker–Planck–Kolmogorov equations with application to periodic numerical homogenization.SIAM J. Numer. Anal., in press

  34. [42]

    Sprekeler and H

    T. Sprekeler and H. V. Tran. Optimal convergence rates for elliptic homogenization problems in nondivergence- form: analysis and numerical illustrations.Multiscale Model. Simul., 19(3):1453–1473, 2021

    2021

  35. [43]

    Z. Wang, J. Xin, and Z. Zhang. Computing effective diffusivity of chaotic and stochastic flows using structure- preserving schemes.SIAM J. Numer. Anal., 56(4):2322–2344, 2018

    2018

  36. [44]

    Z. Wang, J. Xin, and Z. Zhang. Sharp error estimates on a stochastic structure-preserving scheme in computing effective diffusivity of 3D chaotic flows.Multiscale Model. Simul., 19(3):1167–1189, 2021

    2021

  37. [45]

    Zhao and E

    L. Zhao and E. Chung. Constraint energy minimizing generalized multiscale finite element method for convection diffusion equation.Multiscale Model. Simul., 21(2):735–752, 2023

    2023

  38. [46]

    K. C. Zygalakis. On the existence and the applications of modified equations for stochastic differential equations. SIAM J. Sci. Comput., 33(1):102–130, 2011. 24 T. SPREKELER, H. WU, AND Z. ZHANG Department of Mathematics, Texas A&M University, College Station, TX 77843, USA. ...

    2011

Pith tools

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