Pith. sign in

REVIEW 1 major objections 4 minor 42 references

Galerkin-collocation approximation in time for the wave equation and its post-processing

T0 review · 1 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read The paper claims that a Galerkin-collocation time-stepping family solves the wave equation at optimal order in space and time while producing C1-in-time discrete solutions.

desk verdict New cGP-C1 time discretization for the wave equation with clean numerics, but the central error proof rests on an unproved identification with [12] and the C2 theorem is deferred; deserves peer review with major revision. read the letter →

arxiv 1908.08238 v1 pith:EPSDB6PC submitted 2019-08-22 math.NA cs.NA

classification math.NAcs.NA MSC 65M6065M1235L05
keywords waveequationGalerkin-collocationmethodcontinuousGalerkin-PetrovcollocationtimediscretizationHermitequadratureerrorestimatespost-processingenergyconservation
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

This paper introduces a family of time-stepping schemes for the wave equation, called cGP-C1(k)-cG(r), that combine the accuracy of continuous Galerkin-Petrov methods with the cheaper algebraic structure of collocation methods. The paper claims that, with a special Hermite-type quadrature and derivative collocation conditions at the time nodes, the fully discrete solution is globally C1 in time and converges at optimal order: $\|e^0\|+\|e^1\| \lesssim \tau^{k+1}+h^{r+1}$ and $\|\nabla e^0\| \lesssim \tau^{k+1}+h^r$ for all times, with the analogous $L^2$-in-time bounds. If true, this gives a way to compute wave propagation with higher time regularity and reduced linear-system cost compared with standard cGP(k) schemes, without losing accuracy. The paper also extends the construction to twice differentiable solutions and shows how a cheap post-processing lifts a cGP-C1(k) solution to a cGP-C2(k+1) solution.

What carries the argument

The mechanism is the matched triple of trial space $P_k$, test space $P_{k-3}$, and the Hermite-type quadrature $Q^H_n$ with derivative weights at the endpoints, together with the Hermite interpolation $I^H_\tau$ of the right-hand side. Collocation conditions at $t^+_{n-1}$ and $t^-_n$ force the discrete solution to satisfy the differential equation pointwise at interval ends, which yields global $C^1$ continuity; Lemma 3.5 then rewrites the scheme with Gauss-Lobatto quadrature and an enlarged test space $P_{k-2}$. The error proof rests on the decomposition $E=\Theta+E_{\tau,h}$, the stability identity of Lemma 5.9, the boundedness estimate of Lemma 5.10, and the consistency-error identity of Lemma 5.7, with the cGP($k-1$) estimates of the companion analysis supplying the core bounds.

What would settle it

Solve the cGP-C1(3)-cG(3) scheme on one time interval and compare it term-by-term with the post-processed cGP(2)-cG(3) solution from the companion analysis using the modified right-hand side $I^H_\tau f$; if the two do not coincide, or if the energy identity of Lemma 5.9 fails by more than round-off, the proof of Theorem 5.13 collapses. In the same computation, check whether the observed temporal order for $\|e^0\|+\|e^1\|$ is exactly $\tau^4$; any systematic order drop falsifies the optimal-order claim.

Watch

Extended reading notes

Core claim

The central claim is that the wave equation can be discretized in time by a Galerkin-collocation scheme whose trial space has polynomial degree $k$, whose test space has degree only $k-3$, and whose endpoint conditions enforce $\partial_t U_{\tau,h}(t_\pm) = -A_h U_{\tau,h} + P_h F$ at the interval ends. Theorem 5.13 states that, for the error $E=(e^0,e^1)=(u-u^0_{\tau,h},\partial_t u-u^1_{\tau,h})$ with initial value $(R_h u_0,R_h u_1)$, the bounds $\|e^0(t)\|+\|e^1(t)\| \lesssim \tau^{k+1}+h^{r+1}$ and $\|\nabla e^0(t)\| \lesssim \tau^{k+1}+h^r$ hold for all $t$, together with the corresponding $L^2(I;H)$ estimates. The proof first bounds the time derivative of the error by viewing $\partial_t U_{\tau,h}$ as a cGP($k-1$) approximation of $\partial_t U$ with a modified right-hand side, then decomposes $U-U_{\tau,h}$ into an interpolation error and a discrete error and combines weak stability, boundedness, and consistency estimates. If correct, the scheme achieves the same order as the standard cGP(k) method while building C1 regularity into the discrete solution itself, and it conserves the discrete energy $\|u^1_{\tau,h}(t_n)\|^2+\|\nabla u^0_{\tau,h}(t_n)\|^2$ at time nodes when $f\equiv 0$.

Load-bearing premise

The analysis assumes, without proving it, that the new cGP-C1(k) solution is exactly the post-processed cGP(k-1) solution of the companion analysis, with polynomial degree $k$ matching $k+1$ there; if that identification fails, the imported stability, boundedness, and initial-value lemmas do not apply and Theorem 5.13 does not follow, and the C2 post-processing theorem likewise rests on a proof listed as in preparation.

Editorial extensions

If this is right

  • The cGP-C1(k)-cG(r) solution is globally once continuously differentiable in time by construction, so no post-processing is needed for C1 regularity.
  • For $k=3$ the test space shrinks to piecewise constants in time, which simplifies the block structure of the linear systems and is the advertised computational saving over standard cGP(k).
  • For vanishing source $f\equiv 0$, the scheme conserves the discrete energy $\|u^1_{\tau,h}(t_n)\|^2+\|\nabla u^0_{\tau,h}(t_n)\|^2$ exactly at every time node.
  • Applying the post-processing of Theorem 7.3 to a cGP-C1(k) solution produces a cGP-C2(k+1) solution, and the numerical experiments show the convergence order rising from $k+1$ to $k+2$.
  • The construction transfers to other second-order hyperbolic systems, such as elastic waves and Maxwell-type problems, for which the wave equation serves as the prototype.

Reading between the lines

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

  • If the asserted identification with the post-processed cGP(k-1) solution is proved directly on each time interval, the whole error analysis becomes a corollary of the companion cGP analysis; that is the most direct route to closing the proof gap.
  • The built-in C1 regularity makes these discrete solutions well suited for a posteriori error control and adaptive time refinement, since nodal derivative values are already part of the discrete representation; the paper hints at this but does not develop it.
  • For nonlinear or non-self-adjoint hyperbolic problems the energy conservation lemma will generally fail, but the collocation-plus-reduced-test-space mechanism is a natural candidate for optimal-order convergence if the stability lemma can be adapted.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

1 major / 4 minor

Summary. The paper introduces a family of Galerkin-collocation time discretizations for the linear wave equation, called cGP-C1(k)-cG(r) for k ≥ 3, in which the unknown U_τ,h is globally C1 in time. The scheme is defined by combining endpoint collocation conditions (3.1b)-(3.1c) with a reduced test space P_{k-3} and a Hermite-type quadrature rule. The main analytical claim is Theorem 5.13, which asserts optimal-order error estimates, ‖e0(t)‖ + ‖e1(t)‖ ≲ τ^{k+1} + h^{r+1} and ‖∇e0(t)‖ ≲ τ^{k+1} + h^r, together with their L2-in-time analogues, under an initial value U0,h = (R_h u0, R_h u1). The proof is built on an error decomposition and on stability and boundedness lemmas recalled from the prior paper [12], based on an asserted correspondence between the present cGP-C1(k) solution and the post-processed cGP(k−1) solution of [12]. In addition, Section 7 proposes a C2-in-time variant cGP-C2(k) and a post-processing relation (Theorem 7.3) whose proof is deferred to reference [16]. Numerical experiments in Section 8 confirm the predicted convergence orders for k = 3 and k = 4, including the order-increase after post-processing.

Significance. If the central error estimates are valid, the paper makes a useful contribution: it constructs a time discretization that achieves optimal-order convergence for the wave equation while maintaining global C1 regularity in time by construction, and the numerical experiments substantiate the predicted rates. The paper also proves energy conservation at discrete nodes (Lemma 6.1), which is a meaningful structural property, and it provides a computationally cheap post-processing route from the C1 family to the C2 family. The genuine novelty is the scheme itself and the consistency analysis of Section 5.1 (Theorem 5.1), together with the quadrature-based identifications in Lemmas 3.5 and 3.6. However, the verification of the central claim is not self-contained: the decisive stability and boundedness lemmas (Lemmas 5.9 and 5.10) and the initial-value lemma (Lemma 5.3) are imported from [12] under an unproved correspondence, and the key theorem of the C2 extension (Theorem 7.3) is deferred to an 'in preparation' reference. These gaps must be closed before the error estimates can be considered established.

major comments (1)
  1. [Section 5.1, Lemma 5.3] Lemma 5.3, which identifies the initial time derivative ∂_t U_τ,h(0) with (R_h, P_h)∂_t U(0), is also deferred: the proof is referred to [12, Lemma 5.4] under the same correspondence that is not established. This lemma is used in the proof of Theorem 5.5 to verify the discrete initial condition required by Theorem 5.4. The same identification issue therefore also affects the auxiliary error estimate in Theorem 5.5, on which Lemma 5.12 relies. A proof of Lemma 5.3 for the collocation scheme of Problem 3.3 should be included.
minor comments (4)
  1. [Section 2.1, Assumption 2.2] The regularity assumption is stated as 'we tacitly assume that the solution u ... satisfies all the additional regularity conditions that are required in our analysis', with f ∈ C^s([0,T];H) for 'some sufficiently large parameter s'. The manuscript should state explicitly which Sobolev or C^m norms of u, ∂_t u, and f are needed for Theorems 5.4, 5.5 and 5.13, because the constants in the error estimates depend on these norms.
  2. [Section 7, Problem 7.1] The quadrature formula denoted Q^H_{n,k} in condition (7.1f) is not defined. The text notes that it differs from Q^H_n of (2.11) by using k evaluations of function values, but the definition should be given explicitly, including its exactness properties, before it is used.
  3. [Section 8.2] The paper states that applying the cGP-C2(5) method directly gives 'exactly the same errors' as the post-processed solution ~E, but no data or derivation for this claim is shown. Since the direct method is not implemented in the table, this statement should either be substantiated or removed.
  4. [References] Reference [16] is listed as 'in preparation' and reference [6] as 'submitted'. Since both are used in the main text as sources of proofs or details, the manuscript should provide updated references or enough self-contained material so that the claims do not depend on unpublished work.

Circularity Check

2 steps flagged · score 4.0 of 10

Main error bound is not fitted or manufactured, but it imports decisive stability and boundedness lemmas from self-authored [12] via an unproved identification; the C2 post-processing theorem is deferred to an in-preparation self-citation.

  1. self citation load bearing [Section 5, preamble before Lemma 5.9; Lemmas 5.9, 5.10, and Lemma 5.12]
    "Exploiting the correspondence of Uτ,h in this paper to LτUτ,h in [12] and keeping in mind that k here is related to k + 1 there, we can recall from [12] the results of stability (cf. [12, Lemma 5.10]) and boundedness (cf. [12, Lemma 5.11])."

    Lemma 5.9 (stability) and Lemma 5.10 (boundedness) are not proved in this paper; they are imported from [12], which is co-authored by one of the present authors, on the strength of an asserted identification between the present Uτ,h and [12]'s LτUτ,h with shifted polynomial degree. The preceding paragraph describes this relation only as 'motivated by the fact... could also be interpreted as...', not as a verified equivalence. Lemma 5.12 then says the decisive upper bound 'coincides with that in Eq. (5.46) of [12]' and that 'our Eτ,h can be identified with ~Eτ,h of [12]', and Theorem 5.13 is drawn from Lemma 5.12.

  2. self citation load bearing [Section 7, after Theorem 7.3]
    "We refer to [16] for details on post-processing techniques for general nonlinear systems of ordinary differential equations and the proof of the analogue to Theorem 7.3."

    Theorem 7.3 is the paper's stated connection between the cGP-C1 and cGP-C2 families and is used in Section 8.2 to interpret the Table 8.2 rates as optimal order for cGP-C2(5). No proof is supplied; the proof is deferred to reference [16], which the bibliography lists as 'in preparation' and whose authors include two of the present authors. This makes the C2/post-processing claim depend on an unverifiable self-citation chain. It is not circular in the fit-to-data sense, and it does not affect Theorem 5.13, but it is a load-bearing self-citation for the extension claim.

full rationale

The paper's main cGP-C1 error estimates are not circular in the statistical or definitional sense: no parameter is fitted to the data used in the prediction, and the numerical experiments of Section 8 compare the discrete solution against exact manufactured solutions, confirming the predicted rates externally. Theorem 5.1, Lemma 5.7, and Lemma 5.11 are proved in the paper, and Theorem 5.4 is a cGP convergence result whose origin is external to the present work. However, the proof of the central Theorem 5.13 passes through Lemma 5.12, whose decisive stability and boundedness ingredients are Lemmas 5.9 and 5.10, both recalled from [12] based on an asserted identification of the present scheme with the post-processed solution of [12]. The identification is described only as 'motivated' and 'could also be interpreted as', not proved from Problem 3.3, and Lemma 5.12 explicitly identifies its Eτ,h with the ~Eτ,h of [12]. Thus the central estimate is partly carried by a self-citation chain in which the key transfer step is unverified. Separately, the C2 extension and post-processing claim of Theorem 7.3 is deferred to reference [16], listed as 'in preparation', so that part of the paper is not self-contained. These are load-bearing gaps and self-citations, but the scheme has independent content, the cGP-C1 analysis has new auxiliary results, and the numerical evidence is external, so the overall circularity score is moderate rather than severe.

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

No free parameters are fitted: the method contains no constants tuned to data, and experimental orders are measured rather than fit. The central results rest on standard interpolation and quadrature theory, on regularity assumptions on the exact solution (Assumption 2.2), and on an asserted correspondence with the cGP analysis of [12] that imports stability and boundedness lemmas. The post-processing theorem for the C2 family is assumed from the unpublished reference [16]. No new physical entities are introduced.

assumptions (5)
  • domain assumption The solution of (2.1) possesses the additional temporal and spatial regularity assumed in Assumption 2.2, including f ∈ C^s([0,T];H) for sufficiently large s.
    Invoked throughout Section 5 to justify interpolation error bounds and the application of Theorem 5.4.
  • domain assumption Theorems and lemmas from [12] (Theorem 5.4, Lemmas 5.3, 5.9, 5.10) are valid for the scheme of Problem 3.3 under an asserted correspondence, with k here replacing k+1 in [12].
    Section 5 states the correspondence and imports stability, boundedness, and initial-value results without proving the correspondence.
  • domain assumption Theorem 7.3, stating that post-processing of cGP-C1(k) yields the cGP-C2(k+1) solution, is assumed; its proof is deferred to reference [16] (in preparation).
    Section 7 relies on this theorem for the C2 extension; no proof is given in the current paper.
  • standard math Quadrature exactness of Hermite, Gauss, and Gauss-Lobatto rules for polynomials of degree up to 2k-3, together with the interpolation properties of Lemmas 4.1-4.5.
    Used throughout Sections 3 and 5, for example in Lemma 3.5 and Theorem 5.1.
  • standard math Well-posedness and regularity of the wave equation (Problem 2.1 has a unique solution) as cited from [34,35].
    Provides the background existence, uniqueness, and continuity results for the continuous problem.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Galerkin-collocation approximation in time for the wave equation and its post-processing." pith.science (2026). https://pith.science/paper/EPSDB6PC

@misc{pith2026190808238,
  author       = {Pith},
  title        = {Pith review of: Galerkin-collocation approximation in time for the wave equation and its post-processing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EPSDB6PC}},
  note         = {Machine review of arXiv:1908.08238}
}
read the original abstract

We introduce and analyze a class of Galerkin-collocation discretization schemes in time for the wave equation. Its conceptual basis is the establishment of a direct connection between the Galerkin method for the time discretization and the classical collocation methods, with the perspective of achieving the accuracy of the former with reduced computational costs provided by the latter in terms of less complex linear algebraic systems. Continuously differentiable in time discrete solutions are obtained by the application of a special quadrature rule involving derivatives. Optimal order error estimates are proved for fully discrete approximations based on the Galerkin-collocation approach. Further, the concept of Galerkin-collocation approximation is extended to twice continuously differentiable in time discrete solutions. A direct connection between the two families by a computationally cheap post-processing is presented. The error estimates are illustrated by numerical experiments.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 42 canonical work pages

  1. [12]

    Post-processed Galerkin approximation of improved order for wave equations

    M. Bause, U. K¨ ocher, F. A. Radu, F. Schieweck, Post-processed Galerkin approxi- mation of improved order for wave equations , Math. Comp., in press (2018), DOI: https://doi.org/10.1090/mcom/3464, pp. 1–34; arXiv:1803.03005

  2. [17]

    Becher, G

    S. Becher, G. Matthies, D. Wenzel, Variational methods for stable time discretization of first -order differential equations , in K. Georgiev, M. Todorov M, I. Georgiev (eds), Advanced Computing in Industrial Mathematics. BGSIAM , Springer, Cham, 2018, pp. 63–75

  3. [16]

    Becher, G

    S. Becher, G. Matthies, Variational time discretizations of higher order and highe r regularity , in preparation

  4. [1]

    Ahmed, S

    N. Ahmed, S. Becher, G. Matthies, Higher-order discontinuous Galerkin time stepping and loc al projection stabilization techniques for the transient Sto kes problem, Comp. Meth. Appl. Mech. Eng., 313 (2017), pp. 28–52

  5. [2]

    Ahmed, G

    N. Ahmed, G. Matthies, L. Tobiska, H. Xie, Discontinuous Galerkin time stepping with local projec- tion stabilization for transient convection–diffusion-re action problems , Comp. Meth. Appl. Mech. Eng., 200 (2011), pp. 1747–1756. 22

  6. [3]

    Ainsworth, P

    M. Ainsworth, P. Monk, W. Muniz, Dispersive and dissipative properties of discontinuous Ga lerkin finite element methods for the second-order wave equation , J. Sci. Comput., 27 (2006), pp. 5–40

  7. [4]

    Akrivis, C

    G. Akrivis, C. Makridakis, R. H. Nochetto, Optimal order a posteriori error estimates for a class of Runge–Kutta and Galerkin methods , Numer. Math., 114 (2009), pp. 133–160

  8. [5]

    Akrivis, C

    G. Akrivis, C. Makridakis, R. H. Nochetto, Galerkin and Runge–Kutta methods: unified formu- lation, a posteriori error estimates and nodal superconver gence, Numer. Math., 118 (2011), pp. 429–456

Show all 42 references
  1. [6]

    Anselmann, M

    M. Anselmann, M. Bause, Numerical study of Galerkin–collocation approximation in time for the wave equation , Proceedings of the Conference on Mathematics of Wave Phenome na, submitted (2019), pp. 1–20; arXiv:1905.00606

  2. [7]

    A. K. Aziz, P. Monk, Continuous finite elements in space and time for the heat equa tion, Math. Comp., 52 (1989), pp. 255–274

  3. [8]

    Bales, I

    L. Bales, I. Lasiecka, Continuous finite elements in space and time for the nonhomog eneous wave equation, Comput. Math. Appl., 27 (1994), pp. 91–102

  4. [9]

    Bangerth, T

    W. Bangerth, T. Heister, G. Kanschat, deal.II, Differential equations analysis library , Technical Reference, http://www.dealii.org, 2013

  5. [10]

    Bangerth, M

    W. Bangerth, M. Geiger, R. Rannacher, Adaptive Galerkin finite element methods for the wave equation, Comput. Meth. Appl. Math., 10 (2010), pp. 3–48

  6. [11]

    J. D. De Basabe, M. K. Sen, M. F. Wheeler, The interior penalty discontinuous Galerkin method for elastic wave propagation: grid dispersion , Geophys. J. Int., 175 (2008), pp. 83–95

  7. [13]

    Bause, U

    M. Bause, U. K¨ ocher, Variational time discretization for mixed finite element ap proximations of nonstationary diffusion problems , J. Comput. Appl. Math., 289 (2015), pp. 208–224

  8. [14]

    Bause, F

    M. Bause, F. A. Radu, U. K¨ ocher, Space-time finite element approximation of the Biot poroela sticity system with iterative coupling , Comput. Methods Appl. Mech. Engrg., 320 (2017), pp. 745–768

  9. [15]

    Bause, F

    M. Bause, F. A. Radu, U. K¨ ocher, Error analysis for discretizations of parabolic problems u sing continuous finite elements in time and mixed finite elements i n space, Numer. Math., 137 (2017), pp. 773–818

  10. [18]

    ˇCesenek, M

    J. ˇCesenek, M. Feistauer, Theory of the space-time discontinuous Galerkin method for nonstationary parabolic problems with nonlinear convection and diffusion , SIAM J. Numer. Anal., 50 (2012), pp. 1181–1206

  11. [19]

    L. F. Demkowicz, J. Gopalakrishnan, An overview of the discontinuous Petrov–Galerkin method , in X. Feng, O. Karakashian, Y. Xing (eds.), Recent Developments in Discontinuous Galerkin Finite Element Methods for Partial Differential Equations , Springer, Cham, 2014, pp. 149–180

  12. [20]

    D¨ orfler, S

    W. D¨ orfler, S. Findeisen, C. Wieners, Space-time discontinuous Galerkin discretizations for li near first-order hyperbolic evolution systems , Comput. Methods Appl. Math., 16 (2016), pp. 409–428

  13. [21]

    A. Ern, F. Schieweck, Discontinuous Galerkin method in time combined with a stabi lized finite element method in space for linear first-order pdes , Math. Comp., 85 (2016), pp. 2099–2129

  14. [22]

    L. C. Evans, Partial Differential Equations , American Mathematical Society, Providence, Rhode 23 Island, 2010

  15. [23]

    S. M. Findeisen, A Parallel and Adaptive Space-Time Method for Maxwell’s Equ ations, PhD Thesis, DOI: 10.5445/IR/1000056876, KIT, Karlruhe, 2016

  16. [24]

    D. A. French, T. E. Peterson, A continuous space-time finite element method for the wave eq uation, Math. Comp., 65 (1996), pp. 491–506

  17. [25]

    M. J. Gander, M. Neum¨ uller, Analysis of a new space-time parallel multigrid algorithm f or parabolic problems, SIAM J. Sci. Comput., 38 (2016), A2173–A2208

  18. [26]

    Hussain, F

    S. Hussain, F. Schieweck, S. Turek, Higher order Galerkin time discretizations and fast multig rid solvers for the heat equation , J. Numer. Math., 19 (2011), pp. 41–61

  19. [27]

    Hussain, F

    S. Hussain, F. Schieweck, S. Turek, An efficient and stable finite element solver of higher order in space and time for nonstationary incompressible flow , Internat. J. Numer. Methods Fluids, 73 (2013), pp. 927–952

  20. [28]

    Joulak, B

    H. Joulak, B. Beckermann, On Gautschi’s conjecture for generalized Gauss–Radau and G auss– Lobatto formulae, J. Comp. Appl. Math., 233 (2009), pp. 768–774

  21. [29]

    Johnson, Discontinuous Galerkin finite element methods for second or der hyperbolic problems , Comput

    C. Johnson, Discontinuous Galerkin finite element methods for second or der hyperbolic problems , Comput. Methods Appl. Mech. Engrg., 107 (1993), pp. 117–129

  22. [30]

    Karakashian, C

    O. Karakashian, C. Makridakis, Convergence of a continuous Galerkin method with mesh modifi - cation for nonlinear wave equations , Math. Comp., 74 (2004), pp. 85–102

  23. [31]

    K¨ ocher, M

    U. K¨ ocher, M. Bause, Variational space-time methods for the wave equation , J. Sci. Comput., 61 (2014), pp. 424–453

  24. [32]

    U. K¨ ocher,Variational space-time methods for the elastic wave equati on and the diffusion equation , PhD Thesis, http://edoc.sub.uni-hamburg.de/hsu/volltexte/2015 /3112/, Helmut Schmidt Univer- sity, Hamburg, 2015

  25. [33]

    Langer, S

    U. Langer, S. Moore, M. Neum¨ uller, Space-time isogeometric analysis of parabolic evoluation equa- tions, Comput. Methods Appl. Mech. Engrg., 306 (2016), pp. 342–363

  26. [34]

    Lions, Optimal Control of Systems Governed by Partial Differential Equations, Springer, Berlin, 1971

    L. Lions, Optimal Control of Systems Governed by Partial Differential Equations, Springer, Berlin, 1971

  27. [35]

    Lions, E

    L. Lions, E. Magenes, Non-Homogeneous Boundary Value Problems and Applications , Springer, Berlin, 1972

  28. [36]

    Makridakis, R

    C. Makridakis, R. H. Nochetto, A posteriori error analysis for higher order dissipative me thods for evolution problems, Numer. Math., 104 (2006), pp. 489–514

  29. [37]

    Matthies, F

    G. Matthies, F. Schieweck, Higher order variational time discretizations for nonline ar systems of ordinary differential equations , Preprint 23/2011, Fakult¨ at f¨ ur Mathematik, Otto-von-Guericke- Universit¨ at Magdeburg, 2011

  30. [38]

    Mikeli´ c, M

    A. Mikeli´ c, M. F. Wheeler, Theory of the dynamic Biot–Allard equations and their link t o the quasi-static Biot system , J. Math. Phys., 53 (2012), pp. 123702:1–15

  31. [39]

    Quarteroni, A

    A. Quarteroni, A. Valli, Numerical Approximation of Partial Differential Equations , Springer, Berlin, 2008

  32. [40]

    Steinbach, Space-time finite element methods for parabolic problems , Comput

    O. Steinbach, Space-time finite element methods for parabolic problems , Comput. Meth. Appl. Math., 15 (2015), pp. 551–566

  33. [41]

    Steinbach, H

    O. Steinbach, H. Yang, Comparison of algebraic multigrid methods for an adaptive s pace-time finite-element discretization of the heat equation in 3D and 4D, Numer. Linear Algebra Appl., 25 (2018), e2143

  34. [42]

    Thome´ e,Galerkin Finite Element Methods for Parabolic Problems , Springer, Berlin, 2006

    V. Thome´ e,Galerkin Finite Element Methods for Parabolic Problems , Springer, Berlin, 2006. 24

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.