Pith. sign in

REVIEW 4 major objections 3 minor 1 cited by

A posteriori error estimates for the wave equation with mesh change in the leapfrog method

T0 review · 4 major / 3 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read The paper proves that for the leapfrog method with time-varying meshes and local time stepping, the full energy-norm error is bounded by a sum of computable indicators, allowing certified adaptive explicit wave propagation.

desk verdict The right target and a mostly sound machinery, but the main theorem has a genuine hole at the initial and final time layers that needs fixing before the fully computable claim stands. read the letter →

arxiv 2411.16933 v2 pith:QBRTS3RC submitted 2024-11-25 math.NA cs.NA

Marcus J. Grote , Omar Lakkis , Carina Santos This is my paper · ORCID
classification math.NAcs.NA MSC 65M1565M6065M5035L05
keywords aposteriorierrorestimateswaveequationleapfrogmethodlocaltimesteppingmeshchangeellipticreconstructionenergynormfiniteelements
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 proves that the error of an explicit leapfrog finite element solution of the wave equation can be bounded a posteriori by a sum of computable indicators, even when the spatial mesh changes at every time step and when finer regions take smaller local time steps. The bound controls the energy norm of the displacement error and the $L^2$ norm of the velocity error at all discrete times. Each contribution to the error—spatial discretization, time discretization, mesh transfer, local time stepping, data approximation, and initial data—is isolated into an indicator that can be evaluated from the computed solution alone. If correct, this removes the last obstacle to fully adaptive, explicit, and certified wave propagation solvers.

What carries the argument

The central objects are the elliptic reconstructors $\omega^n = R_n U^n$ and $\psi^{n-1/2} = R_n V^{n-1/2}$, which map the discrete displacement and velocity into the continuous space $V$ by inverting the discrete elliptic operator. The argument then builds piecewise linear and quadratic time interpolants on the primal and staggered time grids, so that the reconstruction-exact error $\sigma = (\breve\omega - u,\breve\psi - v)$ solves the first-order error-residual system (4.5). The energy estimate (4.7) bounds $\|\sigma\|_{L^\infty(0,T;\mathrm{erg},A)}$ by the initial error plus twice the $L^1$ energy norm of the residuals, and each residual term is further decomposed using the mesh-change, local-time-stepping, time, data, and elliptic indicators of Section 4.3.

What would settle it

Take the same leapfrog scheme on a time-varying mesh that is deliberately not compatible in the sense of Appendix A.1 (for example, refine one half of the domain and coarsen the other independently at a single time step), compute both the true maximum energy error and the right-hand side of (4.19), and check whether the bound is violated; a violation would show that the compatibility condition is essential rather than an artifact of the proof.

Watch

Extended reading notes

Core claim

The central claim is Theorem 4.4: for the fully discrete leapfrog scheme (2.55) on time-varying finite element spaces, the maximum over all time steps of the energy-norm displacement error and of the $L^2$ velocity error is bounded by the sum of a computable residual estimator, the initial error, and a time-accumulated combination of elliptic, mesh-change, local-time-stepping, time-discretization, and data-approximation indicators. The proof introduces piecewise linear and quadratic reconstructions in time, transfers the discrete solution between consecutive meshes with abstract grid transfer operators, and bounds the resulting residuals by indicators that are fully computable. Numerical experiments with a moving Gaussian pulse confirm that the estimator converges at the optimal rate with mesh refinement.

Load-bearing premise

The proof assumes every spatial mesh in the time sequence is derived from one common macro triangulation by bisection, so consecutive meshes are always compatible; the Clément-Scott-Zhang and transfer-operator bounds that make the mesh-change indicators valid depend on this compatibility.

Editorial extensions

If this is right

  • Every term in the bounds (4.19) and (4.20) is fully computable, so an adaptive solver can monitor the actual error estimate during the run and refine or coarsen wherever the local indicators are largest.
  • The analysis covers leapfrog-based local time stepping with two or more local time steps, including stabilized variants, so the CFL restriction can be relaxed locally without losing the error bound.
  • On a fixed mesh the estimates reduce to known a posteriori bounds for leapfrog, recovering the optimal convergence rate $O(h)$ in the energy norm.
  • The numerical example shows that the estimator converges at the expected optimal rate with mesh size even when the mesh moves with the wave front.
  • The bounds account for mesh-change error explicitly, which is the key ingredient for adaptive refinement and coarsening over long integration times.

Reading between the lines

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

  • An adaptive algorithm that uses these indicators can in principle drive the whole space-time mesh hierarchy, but the theory does not address the computational cost of evaluating the indicators themselves; in practice the elliptic estimators may dominate the runtime.
  • The compatibility condition (common macro triangulation) is likely necessary for the Clément-Scott-Zhang estimates used in Lemma A.5, so an arbitrary pair of time-consecutive meshes could make the transfer indicators fail, which is a testable limitation.
  • The same reconstruction strategy may extend to higher-order time-stepping methods or to nonlinear wave equations, but those extensions are not shown here.
  • Since the indicators bound the discrete solution error rather than the energy-conservation defect, comparing the mesh-change indicators with the actual energy jump at each mesh change could reveal how much of the energy drift is captured by the estimator.
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 / 3 minor

Summary. The paper derives a fully computable a posteriori error estimator for the Galerkin finite element solution of the wave equation with explicit leapfrog time-stepping, allowing for time-varying meshes and leapfrog-based local time-stepping. The main results, Theorem 4.4, bound the maximum energy-norm error of the displacement and the L2 error of the velocity by a sum of computable indicators: elliptic indicators, mesh-change indicators, LTS indicators, time-discretization indicators, data-approximation indicators, and an initial-error term. The proof uses elliptic reconstruction, abstract grid transfer operators, and quadratic time reconstructions. Numerical experiments for a one-dimensional Gaussian pulse on a locally refined, time-varying mesh show optimal convergence rates of the estimator with respect to mesh size.

Significance. If correct, the result is significant: it provides the first fully computable a posteriori error bound for an explicit leapfrog finite element wave solver that simultaneously accounts for mesh change and local time-stepping, thereby opening the way to rigorous adaptive explicit wave propagation. The estimator is parameter-free and does not rely on unknown constants, and the numerical experiments confirm the expected convergence rates. The analysis builds on prior work by the authors and collaborators but extends it in a nontrivial way to time-varying conforming meshes and LTS. The abstract transfer-operator framework and the explicit treatment of mesh-change indicators are elegant and likely useful for future adaptive algorithms. However, the proof as written has gaps at the initial and final time layers that directly affect the advertised fully computable bound, so the central claim is not yet fully established.

major comments (4)
  1. [§3.3–§4.4, Eqs. (3.5), (3.6), (4.26)–(4.33)] The residuals ρ_n^0 and ρ_1^{n-1/2} are defined in (3.5) only for n = 1, ..., N−1, yet the piecewise constant extensions (3.6) sum over n = 0, ..., N. Consequently the error-residual PDE (4.5) and the bounds (4.28) and (4.32) are used on the first and last half-intervals I_{1/2} and I_{N+1/2}, where they require ρ_0^0, ρ_0^N, ρ_1^{-1/2} and ρ_1^{N-1/2}, together with quantities such as U^{-1}, ω^{-1/2}, V^{N+1/2} and ω^{N+1/2} that the scheme (2.54) never defines. The integration of ||r|| over [0,T] in (4.33) therefore is not justified as written. The authors must either define the missing ghost values and endpoint residuals explicitly, or restrict the theorem to a setting where the first and last half-intervals are handled by a separate argument.
  2. [§4.4, Eq. (4.34)] The assertion ||σ(0)||_{erg,A} ≤ ||e(0)||_{erg,A} is stated without proof and is not a consequence of the discrete initial data being Ritz/L2 projections. Indeed σ_1(0) = ˘ψ(0) − v(0), and (3.11) defines ˘ψ(0) through an integral of Aω̂ over [t_{−1/2}, 0] plus the term (Δt/2)(R_1Π_1F^0 + ρ_0^0); this generally contains contributions proportional to Δt A_0U^0 (or to the undefined ω^{−1/2}) that are not present in e(0) if e(0) is taken as the projection error (U^0 − u(0), V^{−1/2} − v(0)). If e(0) is meant to include elliptic reconstruction errors, then the estimator should also include ε_0^0 and ε_1^0 with n = 0 in the maxima of (4.19)–(4.20), which it does not. Thus the bound (4.34) needs a derivation or the statement of Theorem 4.4 must be amended with an additional initial-layer indicator.
  3. [§4.3, Eq. (4.13) and §4.4, Eq. (4.27)] The mesh-change indicator µ_n^0 is defined using the intersection V_n ∩ V_{n+1}, but the transfer term it bounds in (4.27) is [R_nΠ_n − R_{n−1}]U^{n−1}, which involves the spaces V_{n−1} and V_n. Lemma A.5, which is cited for this bound, gives the estimate in terms of W ∩ V with W and V being the two finite element spaces actually involved; here that is V_{n−1} ∩ V_n. With the definition as written, the bound (4.27) is not established. The authors should correct the intersection in the definition of µ_n^0 (and in any analogous indicators) so that it matches the mesh change from V_{n−1} to V_n.
  4. [§2.10, Eq. (2.55)] The scheme (2.55), stated as the two-step equivalent of the system form (2.54), is inconsistent with (2.54). From (2.54), U^1 = Π^1 U^0 + V^{1/2}Δt = Π^1[U^0 + V^{−1/2}Δt + (F^0 − fA_0U^0)Δt^2] = Π^1[U^0 + P_0v_0Δt + (F^0 − fA_0U^0)Δt^2/2], where the last equality uses the definition of V^{−1/2}. Equation (2.55) instead contains (F^0 − fA_0U^0)Δt^2 without the factor 1/2. Since the two schemes are claimed to be equivalent, this is an error that should be corrected.
minor comments (3)
  1. [§4.3, Eq. (4.18)] The index n = ⌈2m⌉ in the definition of ζ_m appears to be a typo: for m = 1 this gives n = 2 even though the integration interval is [t_0, t_{1/2}], which should be associated with n = 1 (or n = 0 depending on convention). Please check the intended map from m to n.
  2. [§4.3, Eq. (4.15)] The notation 'ℓ_n(t)−1/2' in ϑ_n^0(t) is ambiguous; it should likely be 'ℓ_n(t) − 1/2' or a different expression. Please clarify the intended formula.
  3. [§5.1, Eq. (5.1)] The text refers to a 'Gaussian beam', but the exact solution (5.1) is a one-dimensional Gaussian pulse; consider rewording to avoid confusion with a beam in higher dimensions.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the estimator is computed from the discrete solution and given data; self-citations supply tools, not the target result, and the key reconstruction lemmas are proved in the text.

full rationale

The derivation chain is not circular. The key residual definitions, time reconstructions, and error indicators are all functions of the discrete solution, the data, and the available elliptic estimator E; no fitted parameter is relabelled as a prediction. The self-citations to Lakkis–Makridakis (elliptic reconstruction), Georgoulis–Lakkis–Makridakis–Virtanen (time reconstruction), and Diaz–Grote (local time stepping) provide tools that are either reproduced in the paper (Lemmas 3.5–3.8 and Appendix A) or used as external scheme definitions, and they are not invoked as a uniqueness theorem to force the choice of estimator. The final bound (4.19)–(4.20) also depends on an explicitly stated external input: the elliptic estimator functional E satisfying (3.4); this is an assumption, not a circular reduction. The two caveats identified by the skeptic are correctness gaps, not circularity. First, residuals in (3.5) are declared only for n = 1, …, N−1, yet bounds (4.28) and (4.32) are applied on initial and final half-intervals, where ρ0^0, ρ0^N, and related quantities such as ω^{−1/2} are not defined by the scheme (2.54). Second, (4.34) asserts without proof that ∥σ(0)∥_{erg,A} ≤ ∥e(0)∥_{erg,A}; this does not follow from definitions alone because σ0(0) = R0U0 − u0 differs from e0(0) = U0 − u0 by an elliptic reconstruction error, and σ1(0) involves the quadratic reconstruction on [t_{−1/2}, 0]. These gaps mean the advertised fully computable bound (4.19)–(4.20) is not fully established, but they do not make the result equivalent to its inputs by construction. They should be weighed in a correctness review, not as circularity.

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

The central claim rests on standard finite element theory, an abstract elliptic estimator input, and the compatible-mesh bisection structure. There are no fitted parameters and no new postulated physical or mathematical entities beyond the reconstruction tools, which are established techniques.

assumptions (5)
  • domain assumption There exists a family of a posteriori elliptic estimators E such that ||ω^n - U^n||_Z ≤ E[U^n, V_n, Z] for Z ∈ {V, V', A, L2}.
    Assumed in §3.2 and used throughout §4 to control the elliptic part of the error. The paper gives one concrete realization (Babuška-Rheinboldt) in Appendix A, but the main theorem is stated abstractly.
  • domain assumption Meshes M_{n-1} and M_n are compatible: all meshes are generated by bisection of a common macro triangulation.
    Stated in §2.7 and Appendix A.1; needed for the Clément-Scott-Zhang inequalities and for the transfer operator estimates in Lemmas A.4-A.5 and the mesh-change indicators (4.13).
  • domain assumption The local time-stepping operator has the form (2.45) or is a generic perturbed bilinear form accommodating two or more local steps; the analysis does not cover arbitrary LTS coupling.
    Introduced in §2.8. The estimator bounds depend on the specific structure of the modified operator.
  • standard math Standard Sobolev and finite element theory: Gelfand triple, Lax-Milgram, Scott-Zhang interpolation on compatible meshes.
    Invoked in §2.2 and Appendix A without proof.
  • domain assumption The exact solution and data satisfy the regularity needed for the energy-norm framework; for simplicity homogeneous boundary conditions are assumed, with Remark 2.3 indicating extensions.
    Section 2.3 states the analysis applies under more general conditions but the paper works with simplified hypotheses.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A posteriori error estimates for the wave equation with mesh change in the leapfrog method." pith.science (2026). https://pith.science/paper/QBRTS3RC

@misc{pith2026241116933,
  author       = {Pith},
  title        = {Pith review of: A posteriori error estimates for the wave equation with mesh change in the leapfrog method},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QBRTS3RC}},
  note         = {Machine review of arXiv:2411.16933}
}
read the original abstract

We derive a fully computable aposteriori error estimator for a Galerkin finite element solution of the wave equation with explicit leapfrog time-stepping. Our discrete formulation accommodates both time evolving meshes and leapfrog based local time-stepping (Diaz & Grote, 2009), which overcomes the stringent stability restriction on the time-step due to local mesh refinement. Thus we account for adaptive time-stepping with mesh change in a fully explicit time integration while retaining its efficiency. The error analysis relies on elliptic reconstructors and abstract grid transfer operators, which allows for use-defined elliptic error estimators. Numerical results using the elliptic Babu\v{s}ka-Rheinboldt estimators illustrate the optimal rate of convergence with mesh size of the aposteriori error estimator.

Figures

Figures reproduced from arXiv: 2411.16933 by the authors.

Figure 1
Figure 1. Schematic description of the linear and quadratic time basis functions, ℓν and qν, for some values of ν. Conversely and consistently, given a sequence (ϕ n)n=0,...,N defined on the integer￾time grid (respectively on the staggered time grid ϕ n−1/2  n=0,...,N+1) we will denote by ϕ(t) its continuous piecewise linear interpolation in time whereby ϕ(t) := X ν ϕ ν ℓν(t), i.e., ϕ(t) := ϕ ν−1 ℓν−1(t) +ϕ ν ℓν(t) for tν−1 … view at source ↗
Figure 2
Figure 2. A schematic illustration of the time-reconstructions and cognate time-functions. The values are only for graphing pur￾poses and do not reflect actual ones. ω˘(t) := ω n−1 + Z t tn−1 ψˆ(s) d s + (t − tn−1)ρ n−1/2 1 , t ∈ In ψ˘(t) := ψ n−1/2 − Z t tn−1/2 Aωˆ(s) d s + (t − tn−1/2) (Rn+1Πn+1F n + ρ n 0 ), t ∈ In+1/2 (3.11) 3.5. Lemma (quadratic time-reconstructions interpolate at nodes). The quadratic displacement recon… view at source ↗
Figure 5
Figure 5. Three successive refinements by bisection of a macro element K1 = Cnvx a 1 , b 1 , c 1 [PITH_FULL_IMAGE:figures/full_fig_p025_5.png] view at source ↗

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 variational approach to the analysis of the continuous space-time FEM for the wave equation

    math.NA 2025-01 accept novelty 7.0 of 10

    The French-Peterson space-time continuous FEM for the Hamiltonian wave equation is proved unconditionally stable with T-independent constants, quasi-optimal a priori error rates, and a reliable explicit a posteriori e...

Reference graph

Works this paper leans on

37 extracted references · 23 canonical work pages · cited by 1 Pith paper

  1. [1]

    S. Adjerid. A posteriori finite element error estimation for second-order hyperbolic problems. Comput. Methods Appl. Mech. Engrg., 191 0 (41-42): 0 4699--4719, 2002. ISSN 0045-7825. doi:10.1016/S0045-7825(02)00400-0

    work page doi:10.1016/s0045-7825(02)00400-0 2002

  2. [2]

    S. Adjerid. A posteriori error estimation for the method of lumped masses applied to second-order hyperbolic problems. Comput. Methods Appl. Mech. Engrg., 195 0 (33-36): 0 4203--4219, 2006. ISSN 0045-7825. doi:10.1016/j.cma.2005.08.003

    work page doi:10.1016/j.cma.2005.08.003 2006

  3. [3]

    Ainsworth and J

    M. Ainsworth and J. T. Oden. A posteriori error estimation in finite element analysis. Pure and Applied Mathematics (New York). Wiley-Interscience [John Wiley & Sons], New York, 2000. ISBN 0-471-29411-X. URL http://www.worldcat.org/oclc/61130479

    arXiv 2000

  4. [4]

    Akrivis, C

    G. Akrivis, C. Makridakis, and R. H. Nochetto. A posteriori error estimates for the C rank- N icolson method for parabolic equations. Math. Comp., 75 0 (254): 0 511--531 (electronic), 2006. ISSN 0025-5718. doi:10.1090/S0025-5718-05-01800-4. URL https://doi.org/10.1090/S0025-5718-05-01800-4

    doi:10.1090/s0025-5718-05-01800-4 2006

  5. [5]

    Babu s ka and W

    I. Babu s ka and W. C. Rheinboldt. Error estimates for adaptive finite element computations. SIAM Journal on Numerical Analysis, 15 0 (4): 0 736--754, 1978. ISSN 0036-1429. doi:10.1137/0715049. URL https://mathscinet.ams.org/mathscinet-getitem?mr=483395

    doi:10.1137/0715049 1978

  6. [6]

    Bangerth and R

    W. Bangerth and R. Rannacher. Adaptive finite element techniques for the acoustic wave equation. J. Comput. Acoust., 9 0 (2): 0 575--591, 2001. ISSN 0218-396X. doi:10.1142/S0218396X01000668

    work page doi:10.1142/s0218396x01000668 2001

  7. [7]

    Bangerth, M

    W. Bangerth, M. Grote, and C. Hohenegger. Finite element method for time dependent scattering: Nonreflecting boundary condition, adaptivity, and energy decay. Computer Methods in Applied Mechanics and Engineering, 193 0 (23): 0 2453--2482, 06 2004. ISSN 0045-7825. doi:10.1016/j.cma.2004.01.021

    work page doi:10.1016/j.cma.2004.01.021 2004

  8. [8]

    Bernardi and E

    C. Bernardi and E. S \"u li. Time and space adaptivity for the second-order wave equation. Math. Models Methods Appl. Sci., 15 0 (2): 0 199--225, 2005. ISSN 0218-2025. doi:10.1142/S0218202505000339

    doi:10.1142/s0218202505000339 2005

Show all 37 references
  1. [9]

    D. Braess. Finite elements: theory, fast solvers, and applications in elasticity theory. Cambridge University Press, Cambridge, 3 edition, 04 2007. ISBN 978-0-511-61863-5. doi:10.1017/CBO9780511618635. URL http://www.worldcat.org/oclc/776966850

    2007

  2. [10]

    Braess, V

    D. Braess, V. Pillwein, and J. Sch \"o berl. Equilibrated residual error estimates are p-robust. Computer Methods in Applied Mechanics and Engineering, 198 0 (13-14): 0 1189--1197, 03 2009. ISSN 00457825. doi:10.1016/j.cma.2008.12.010

    2009 doi

  3. [11]

    Carle and M

    C. Carle and M. Hochbruck. Error analysis of multirate leapfrog-type methods for second-order semilinear ODEs . SIAM Journal on Numerical Analysis, 60 0 (5): 0 2897--2924, 10 2022. ISSN 0036-1429. doi:10.1137/21M1427255

    2022 doi

  4. [12]

    Chaumont-Frelet

    T. Chaumont-Frelet . Asymptotically Constant-Free and Polynomial-Degree-Robust a Posteriori Estimates for Space Discretizations of the Wave Equation . SIAM Journal on Scientific Computing, 45 0 (4): 0 A1591--A1620, 08 2023. ISSN 1064-8275. doi:10.1137/22M1485619

    2023 doi

  5. [13]

    Chaumont-Frelet and A

    T. Chaumont-Frelet and A. Ern. Damped energy-norm a posteriori error estimates for fully discrete approximations of the wave equation using C2-reconstructions . Technical Report arXiv:2403.12954, arXiv, 03 2024

    2024 arXiv

  6. [14]

    Chen and F

    Z. Chen and F. Jia. An adaptive finite element algorithm with reliable and efficient error control for linear parabolic problems. Math. Comp., 73 0 (247): 0 1167--1193 (electronic), 2004. ISSN 0025-5718

    2004

  7. [15]

    Cohen, P

    G. Cohen, P. Joly, J. E. Roberts, and N. Tordjman. Higher Order Triangular Finite Elements with Mass Lumping for the Wave Equation . SIAM Journal on Numerical Analysis, 38 0 (6): 0 2047--2078, 01 2001. ISSN 0036-1429. doi:10.1137/S0036142997329554

    2001 doi

  8. [16]

    Diaz and M

    J. Diaz and M. J. Grote. Energy conserving explicit local time stepping for second-order wave equations. SIAM Journal on Scientific Computing, 31 0 (3): 0 1985--2014, 01 2009. ISSN 1064-8275. doi:10.1137/070709414

    1985 doi

  9. [17]

    Diaz and M

    J. Diaz and M. J. Grote. Multi-level explicit local time-stepping methods for second-order wave equations. Computer Methods in Applied Mechanics and Engineering, 291: 0 240--265, July 2015. ISSN 0045-7825. doi:10.1016/j.cma.2015.03.027

    2015 doi

  10. [18]

    T. Dupont. Mesh modification for evolution equations. Math. Comp., 39 0 (159): 0 85--107, 1982. ISSN 0025-5718

    1982

  11. [19]

    Eriksson and C

    K. Eriksson and C. Johnson. Adaptive finite element methods for parabolic problems. I . A linear model problem. SIAM J. Numer. Anal., 28 0 (1): 0 43--77, 1991. ISSN 0036-1429. doi:10.1137/0728003. URL https://doi.org/10.1137/0728003

    1991 doi

  12. [20]

    F. D. Gaspoz, K. Siebert, C. Kreuzer, and D. A. Ziegler. A convergent time--space adaptive dG (s) finite element method for parabolic problems motivated by equal error distribution. IMA Journal of Numerical Analysis, 39 0 (2): 0 650--686, 04 2019. ISSN 0272-4979. doi:10.1093/i...

    2019 doi

  13. [21]

    E. H. Georgoulis, O. Lakkis, C. G. Makridakis, and J. M. Virtanen. A posteriori error estimates for leap-frog and cosine methods for second order evolution problems. SIAM Journal on Numerical Analysis, 54 0 (1): 0 120--136, 01 2016. ISSN 0036-1429. doi:10.1137/140996318. URL h...

    2016 arXiv

  14. [22]

    Gorynina, A

    O. Gorynina, A. Lozinski, and M. Picasso. An easily computable error estimator in space and time for the wave equation. ESAIM: Mathematical Modelling and Numerical Analysis, 53 0 (3): 0 729--747, 05 2019. ISSN 0764-583X, 1290-3841. doi:10.1051/m2an/2018049

    2019

  15. [23]

    Grote, S

    M. Grote, S. Michel, and S. Sauter. Stabilized leapfrog based local time-stepping method for the wave equation. Mathematics of Computation, 90 0 (332): 0 2603--2643, 11 2021. ISSN 0025-5718, 1088-6842. doi:10.1090/mcom/3650

    2021 doi

  16. [24]

    M. J. Grote and T. Mitkova. Explicit local time-stepping methods for Maxwell 's equations. Journal of Computational and Applied Mathematics, 234 0 (12): 0 3283--3302, 10 2010. ISSN 0377-0427. doi:10.1016/j.cam.2010.04.028

    2010 doi

  17. [25]

    M. J. Grote, M. Mehlin, and S. A. Sauter. Convergence analysis of energy conserving explicit local time-stepping methods for the wave equation. SIAM Journal on Numerical Analysis, 56 0 (2): 0 994--1021, 01 2018. ISSN 0036-1429. doi:10.1137/17M1121925

    2018 doi

  18. [26]

    Hairer, C

    E. Hairer, C. Lubich, and G. Wanner. Geometric numerical integration, volume 31 of Springer Series in Computational Mathematics . Springer, Heidelberg, 2010. ISBN 978-3-642-05157-9. URL https://mathscinet.ams.org/mathscinet-getitem?mr=2840298

    2010

  19. [27]

    G. M. Hulbert and T. J. R. Hughes. Space-time finite element methods for second-order hyperbolic equations. Computer Methods in Applied Mechanics and Engineering, 84 0 (3): 0 327--348, 12 1990. ISSN 0045-7825. doi:10.1016/0045-7825(90)90082-W

    1990 doi

  20. [28]

    C. Johnson. Discontinuous G alerkin finite element methods for second order hyperbolic problems. Comput. Methods Appl. Mech. Engrg., 107 0 (1-2): 0 117--129, 1993. ISSN 0045-7825. doi:10.1016/0045-7825(93)90170-3

    1993 doi

  21. [29]

    Karakashian and C

    O. Karakashian and C. Makridakis. Convergence of a continuous G alerkin method with mesh modification for nonlinear wave equations. Math. Comp., 74 0 (249): 0 85--102 (electronic), 2005. ISSN 0025-5718. doi:10.1090/S0025-5718-04-01654-0

    2005 doi

  22. [30]

    Kossaczk \'y

    I. Kossaczk \'y . A recursive approach to local mesh refinement in two and three dimensions. J. Comput. Appl. Math., 55 0 (3): 0 275--288, 1994. ISSN 0377-0427. doi:10.1016/0377-0427(94)90034-5

    1994 doi

  23. [31]

    Lakkis and C

    O. Lakkis and C. Makridakis. Elliptic reconstruction and a posteriori error estimates for fully discrete linear parabolic problems. Math. Comp., 75 0 (256): 0 1627--1658, 2006. ISSN 0025-5718. doi:10.1090/S0025-5718-06-01858-8. URL https://doi.org/10.1090/S0025-5718-06-01858-8

    2006 doi

  24. [32]

    Lakkis and T

    O. Lakkis and T. Pryer. Gradient recovery in adaptive finite-element methods for parabolic problems. IMA J. Numer. Anal., 32 0 (1): 0 246--278, 2012. ISSN 0272-4979. doi:10.1093/imanum/drq019

    2012 doi

  25. [33]

    W. F. Mitchell. A comparison of adaptive refinement techniques for elliptic problems. ACM Trans. Math. Softw., 15 0 (4): 0 326--347, 12 1989. ISSN 0098-3500. doi:10.1145/76909.76912

    1989

  26. [34]

    M. Picasso. Adaptive finite elements for a linear parabolic problem. Comput. Methods Appl. Mech. Engrg., 167 0 (3-4): 0 223--237, 1998. ISSN 0045-7825. doi:10.1016/S0045-7825(98)00121-2. URL http://dx.doi.org/10.1016/S0045-7825(98)00121-2

    1998 doi

  27. [35]

    M. Picasso. Numerical study of an anisotropic error estimator in the l^2(h^1) norm for the finite element discretization of the wave equation. SIAM Journal on Scientific Computing, 32 0 (4): 0 2213--2234, 2010. doi:10.1137/090778249

    2010 doi

  28. [36]

    Schmidt and K

    A. Schmidt and K. G. Siebert. Design of adaptive finite element software, volume 42 of Lecture Notes in Computational Science and Engineering. Springer-Verlag, Berlin, 2005. ISBN 3-540-22842-X. doi:10.1007/b138692. URL http://www.worldcat.org/oclc/488984204. The finite element...

    2005

  29. [37]

    Verf \"u rth

    R. Verf \"u rth. A posteriori error estimation techniques for finite element methods. Numerical Mathematics and Scientific Computation . Oxford University Press, Oxford, 2013. ISBN 978-0-19-967942-3. doi:10.1093/acprof:oso/9780199679423.001.0001. URL http://www.worldcat.org/oc...

    2013

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.