Pith. sign in

REVIEW 2 major objections 4 minor 1 cited by

Inf-sup stable space-time discretization of the wave equation based on a first-order-in-time variational formulation

T0 review · 2 major / 4 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 A conforming space-time discretization of the first-order wave equation with exponential weights is unconditionally stable and quasi-optimal for any standard tensor-product spaces.

desk verdict Solid space-time FEM paper with a genuine new method, but the advertised generality needs an extra regularity assumption that is not stated in the abstract. read the letter →

arxiv 2506.05886 v1 pith:AR7VNWWD submitted 2025-06-06 math.NA cs.NA

Matteo Ferrari , Ilaria Perugia , Enrico Zampa This is my paper
classification math.NAcs.NA MSC 65M6065M1265M1535L05
keywords space-timefiniteelementmethodwaveequationfirst-order-in-timeformulationexponentialweightsunconditionalstabilityinf-supconditionquasi-optimalconvergencetensorproductapproximation
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 proposes a conforming space–time finite-element method for the wave equation written as a first-order-in-time system, with the wave field and its velocity as unknowns. The central claim is that adding exponential weights $e^{-t/T}$ to the time integrals and testing with time derivatives of the trial functions makes the discrete bilinear form inf-sup stable with a constant independent of the mesh, so the method is unconditionally stable for any conforming tensor-product space–time spaces and converges quasi-optimally under standard approximation assumptions. Concretely, the paper proves error bounds of order $h_t^{s_t}+h_x^{s_x}$ in the energy-type norm and $h_t^{s_t+1}+h_x^{s_x+1}$ in the $L^2$ norm under elliptic regularity. A reader should care because most space–time wave-equation methods either demand extra regularity of the solution or special temporal spaces, while this one works with standard polynomial or spline spaces and imposes no CFL-type time-step restriction.

What carries the argument

The carrying object is the exponentially weighted scalar product $(u,v)_{L^2_e(0,T)} = \int_0^T u(s)v(s)e^{-s/T}\,ds$ and the bilinear form $A((U,V),(\lambda,\chi)) = (\partial_t V,\lambda)_{L^2_e(Q_T)} + (c^2\nabla_x U,\nabla_x\lambda)_{L^2_e(Q_T)} - (\partial_t U,\chi)_{L^2_e(Q_T)} + (V,\chi)_{L^2_e(Q_T)}$ on the trial space $H^1_{0,\bullet}(0,T;H^1_0(\Omega))\times H^1_{0,\bullet}(0,T;L^2(\Omega))$. The weight makes integration by parts in time produce the positive boundary term $\frac{1}{2e}|w(T)|^2$ and the positive bulk term $\frac{1}{2T}\|w\|^2$, converting the antisymmetric part of the form into the lower bound that closes the inf-sup argument. The Newton potential $N_\Omega$, the inverse of the operator $-\nabla\cdot(c^2\nabla\cdot)$ with homogeneous Dirichlet conditions, supplies the duality norm $\|\partial_t V\|_{N_{\Omega e}}$ and makes the test function $\lambda$ admissible; discretely, a discrete Newton potential $N^{\Omega}_{h_x}$ plays the same role. Finally, commuting elliptic projectors $\Pi^{\nabla}_{h_x}$ and $\Pi^{\partial_t}_{h_t}$ separate the error into spatial and temporal approximation terms.

What would settle it

Run the scheme (23) on the travelling-wave example (47) with $c=1$ and refine $h_t=h_x$: there $\nabla\cdot(c^2\nabla U)=U_{xx}$ carries a Dirac term across the characteristic $x-t+1=0$, so hypothesis (33) is violated, and the measured energy-norm rate shows directly whether the proof's extra regularity is necessary — a drop below $h_t^{s_t}+h_x^{s_x}$ would falsify the claim that the stated rates follow from the stated assumptions.

Watch

Extended reading notes

Core claim

On its own terms, the discovery is that a first-order-in-time variational formulation of the wave equation can be made unconditionally stable and quasi-optimal for generic conforming tensor-product spaces by keeping exponential weights in the method, not just in the analysis. The engine is the identity $(w,\partial_t w)_{L^2_e(0,T)} = \frac{1}{2T}\|w\|^2_{L^2_e} + \frac{1}{2e}|w(T)|^2$, which turns the skew-symmetric coupling between $U$ and $V$ into positive energy contributions. With the test pair $\lambda = N_\Omega \partial_t V + 2T\partial_t U$, $\chi = -\partial_t U + 2T\partial_t V$, where $N_\Omega$ is the Newton potential, the bilinear form satisfies the inf-sup condition; the same construction works discretely with a discrete Newton potential, yielding stability and well-posedness. The error analysis then combines commuting elliptic projectors with the discrete inf-sup condition to obtain quasi-optimal rates in the energy norm and, under an elliptic-regularity assumption, in the $L^2$ norm.

Load-bearing premise

The proof's load-bearing premise is that $\nabla\cdot(c^2\nabla U)$ has one time derivative in $L^2$, which is stronger than the solution regularity stated in Assumption 2.1 and is used to move the time projection onto the divergence term; if this extra smoothness is absent, the advertised quasi-optimal rate is not proven.

Editorial extensions

If this is right

  • For any conforming tensor-product spaces satisfying standard approximation assumptions, the discrete solution is stable for arbitrary time-step size relative to spatial mesh size; no CFL condition is needed.
  • The same formulation covers continuous piecewise polynomials, maximal-regularity splines, and $C^1$ splines, with no inverse estimates or quasi-uniform meshes required.
  • The energy-norm error is $O(h_t^{s_t}+h_x^{s_x})$; with elliptic regularity the $L^2$ error improves to $O(h_t^{s_t+1}+h_x^{s_x+1})$.
  • The singular test with a traveling wave that violates the extra regularity still shows convergence, so the method appears more robust than the theory in the paper guarantees.

Reading between the lines

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

  • The exponential weight acts as a built-in stabilizer whose effect is independent of the mesh; this suggests trying smaller weights or weights localized near the final time, which might preserve the energy estimates while reducing distortion of the variational problem.
  • The proof structure is not tied to a particular time-stepping space, so it may transfer to other first-order hyperbolic systems with a similar antisymmetric structure, such as variable-coefficient acoustics or Maxwell-type evolutions, where the Newton potential supplies the needed duality norm.
  • The singular numerical test indicates that hypothesis (33) may be an artifact of the proof technique; a convergence analysis that avoids $\nabla\cdot(c^2\nabla U)\in H^1(0,T;L^2)$ would be the natural way to close the gap between theory and numerics.
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

2 major / 4 minor

Summary. The paper introduces a conforming space-time discretization of the first-order-in-time wave equation based on a variational formulation with exponential weights. The main theoretical contributions are: (i) a continuous inf-sup condition in weighted norms built with a Newton potential; (ii) a discrete inf-sup condition that transfers verbatim to any conforming tensor-product spaces, yielding unconditional stability in the discrete energy norm; (iii) quasi-optimal error estimates in the energy norm and in L2, under standard approximation assumptions on the discrete spaces; and (iv) numerical evidence for smooth and singular solutions.

Significance. If the stated scope is accurate, the method is valuable: it gives a unified framework for conforming space-time discretizations of the wave equation with no CFL-type restriction and with rates that match the approximation power of the discrete spaces. The proof strategy is self-contained, uses a clean Newton-potential argument, and the paper includes reproducible numerical experiments with code referenced. The analysis does not fit parameters and does not rely on unproved regularity assumptions beyond those explicitly stated, which is a strength.

major comments (2)
  1. [§4.2, Lemma 4.7 and Theorem 4.15] The quasi-optimal error analysis requires the extra regularity ∇x·(c²∇xU) ∈ H¹(0,T;L²(Ω)) stated in (33). This assumption is used crucially in the estimate of I2, where integration by parts in space and commutation with the elliptic time projector Π∂t_ht are applied to ∇x·(c²∇x(Π∂t_ht−Id)U). Assumption 2.1 and Proposition 2.2 only yield ∂t²U ∈ L∞(0,T;L²), hence ∇x·(c²∇xU) = ∂t²U − F ∈ L∞(0,T;L²), which is not in the domain H¹(0,T;L²) of Π∂t_ht. Thus Theorem 4.15 is not proven for all solutions satisfying the minimal well-posedness assumptions. The abstract's claim of 'quasi-optimal convergence ... under standard approximation assumptions' is therefore broader than the proven statement. The authors should either extend the error analysis to the minimal-regularity setting or explicitly restrict the abstract and introduction, flagging (33) as an additional regularity requirement on the solution.
  2. [§4.3.2] The singular test is outside the scope of Theorem 4.15 precisely because the extra regularity (33) and the other regularity hypotheses on V and on ∇x·(c²∇xU) fail for the traveling wave (47). The numerical convergence O(h^{3/2}) for U and O(h^{1/2}) for V is presented without linking it to the theory. This is not an error, but the paper should state more clearly that the singular test is an empirical phenomenon not covered by the analysis, and should not be read as evidence for the general quasi-optimality claim without the additional regularity.
minor comments (4)
  1. [Corollary 4.3] The stability estimate omits the √T factors that appear in the analogous continuous estimate in Theorem 3.7. Unless the norms of the initial data are understood in the weighted L2e sense, the displayed inequality should read ∥∇x·(c²∇xU0)∥ and ∥c∇xV0∥ multiplied by √T (or the weights should be made explicit). This is a local correction and does not affect the qualitative unconditional-stability claim.
  2. [Proposition 3.3, proof of (18)] The proof of the bound on ∥(λ,χ)∥_We omits the intermediate step in which the cross terms cancel: (NΩ∂tV,∂tU)_L2e equals (∂tU,∂tV)_NΩe, so the +4T(...) term in ∥λ∥²_L2e and the −4T(...) term in ∥χ∥²_NΩe cancel exactly. Adding this one line would prevent the reader from thinking that a positive cross term is being dropped.
  3. [Section 1] The first paragraph contains a duplicated word: 'key properties properties'. This is a typographical error.
  4. [Theorem 4.15] The theorem statement lists the regularity assumptions on U, V, and ∇x·(c²∇xU) immediately after the data assumptions. It would be clearer to label (33) explicitly as an additional regularity assumption on the solution, rather than presenting it as part of the standard approximation framework.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the inf-sup and error estimates are proven from the weighted formulation with explicit test functions and standard approximation assumptions, and the self-citations are contextual only.

full rationale

The paper derives stability and convergence from first principles rather than from fitted parameters or self-referential premises. Proposition 3.3 proves the continuous inf-sup condition by explicitly constructing the test pair in (20) and bounding the resulting norms; Proposition 4.2 repeats the same construction in discrete spaces, so Corollary 4.3 follows without assuming the conclusion. The error analysis in Lemma 4.7 uses Galerkin orthogonality plus the projectors defined in (27)-(28), and Theorems 4.15 and 4.20 combine that with approximation properties under Assumptions 4.8 and 4.9; the rates are derived, not inserted. The extra regularity hypothesis (33) is an additional regularity condition on the exact solution, not an input that is later renamed as a prediction, so it is a generality limitation rather than a circular step. The numerical section explicitly states that the singular test in Section 4.3.2 falls outside the assumptions of Theorem 4.15, which is an honest limitation. Self-citations [12,13,15] appear only in the introduction as related-work comparisons and in the numerical section as an implementation pointer, and they are not used to establish inf-sup stability, quasi-optimality, or any central theorem. There are no fitted constants, no empirical predictions derived from fitted values, and no uniqueness or ansatz imported from the authors' prior work. The central claims are therefore self-contained mathematical results; any concerns about regularity are correctness/generality issues, not circularity.

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

The paper introduces no new physical entities or fitted constants. The free-parameter list is empty because the method has no empirical parameters. The axioms are standard regularity and approximation assumptions, plus the classical existence theory for the wave equation. The additional regularity (33) is the most restrictive hypothesis and is load-bearing for the convergence proof.

assumptions (6)
  • standard math Classical existence and regularity theory for the wave equation (Proposition 2.2), cited from [1, Prop. 4.4], [19, Lemma 2.7], and [10, Chapter 7.2].
    Used to guarantee existence of a solution with sufficient regularity for the variational formulation in Theorem 3.7.
  • domain assumption Assumption 2.1: data regularity, c in L^infty(Omega), c >= c0 > 0, U0 in H^1_0(Omega), V0 in H^1_0(Omega), F in H^1(0,T; L^2(Omega)), F(0) in L^2, div(c^2 grad U0) in L^2.
    Postulated to ensure improved regularity of the solution, needed for the inf-sup analysis and error estimates.
  • domain assumption Assumptions 4.8 and 4.9: approximation properties of the discrete spaces Shx and Sht.
    Standard assumptions on discrete spaces to obtain convergence rates in the error estimates.
  • domain assumption Assumption 4.17: elliptic regularity of the spatial operator.
    Needed only for the improved L2 error estimates in Theorem 4.20.
  • standard math Aubin-Nitsche trick in Lemma 4.11, stated as standard with the proof omitted.
    Used to derive L2 error estimates from H1-type estimates; the argument is standard in finite element theory.
  • domain assumption Additional regularity (33): div_x(c^2 grad_x U) in H^1(0,T; L^2(Omega)).
    Hypothesis of Lemma 4.7, used to integrate by parts in the error analysis. It is not implied by Assumption 2.1.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Inf-sup stable space-time discretization of the wave equation based on a first-order-in-time variational formulation." pith.science (2026). https://pith.science/paper/AR7VNWWD

@misc{pith2026250605886,
  author       = {Pith},
  title        = {Pith review of: Inf-sup stable space-time discretization of the wave equation based on a first-order-in-time variational formulation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AR7VNWWD}},
  note         = {Machine review of arXiv:2506.05886}
}
read the original abstract

In this paper, we present a conforming space-time discretization of the wave equation based on a first-order-in-time variational formulation with exponential weights in time. We analyze the method, showing its stability without imposing any restrictions on the mesh size or time step, and proving quasi-optimal convergence for any choice of space-time tensor product discrete spaces that satisfies standard approximation assumptions. Numerical examples are provided to support the theoretical findings.

Figures

Figures reproduced from arXiv: 2506.05886 by the authors.

Figure 1
Figure 1. Test with the smooth solution given by (46). Relative errors in the norm Ve,h(QT ) for the discrete solution computed with the scheme in (23). The results are obtained by fixing the temporal mesh size ht, decreasing the spatial mesh size hx and varying the polynomial degree p: p = 1, p = 2, p = 3, p = 4, p = 5. The left plot corresponds to maximal regularity splines, while the right plot shows results obtained using… view at source ↗
Figure 2
Figure 2. Test with the smooth solution given by (46). Relative errors in the norm Ve,h(QT ) for the discrete solution computed with the scheme in (23). The results are obtained by varying the spatial and temporal mesh sizes ht = hx and varying the polynomial degree p. The left plot corresponds to maximal regularity splines, while the right plot shows results obtained using C 1 -splines. 0.05 0.07 0.09 0.11 10-10 10-8 10-6 10… view at source ↗
Figure 3
Figure 3. Test with the smooth solution given by (46). Relative errors in the norm L 2 (QT ) for Uh, the first component of the discrete solution computed with the scheme in (23). The results are obtained by varying the spatial and temporal mesh sizes ht = hx and varying the polynomial degree p. The left plot corresponds to maximal regularity splines, while the right plot shows results obtained using C 1 -splines. 21 [PITH_F… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Test with the singular solution given by (47). Relative errors in the norm L 2 (QT ) for Uh (left plot) and Vh (right plot), the two components of the discrete solution with the scheme in (23). Maximal regularity splines are employed in both space and time. The results…

Discussion (0). Sign in 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. Paving the way to a $\operatorname{T}$-coercive method for the wave equation

    math.NA 2025-09 conditional novelty 6.0 of 10

    A parameter-dependent transformation T_mu makes the weak form of u''+mu u=f coercive with mu-independent stability, a step toward a T-coercive space-time wave equation method.

Reference graph

Works this paper leans on

30 extracted references · 24 canonical work pages · cited by 1 Pith paper

  1. [1]

    Abdulle and P

    A. Abdulle and P. Henning. Localized orthogonal decomposition method for the wave equation with a continuum of scales. Math. Comp. , 86(304):549–587, 2017

    work page 2017

  2. [2]

    Anselmann and M

    M. Anselmann and M. Bause. Numerical Study of Galerkin–Collocation Approxima- tion in Time for the Wave Equation. In Mathematics of Wave Phenomena , pages 15–36, Cham, 2020. Springer

    work page 2020

  3. [3]

    Anselmann, M

    M. Anselmann, M. Bause, S. Becher, and G. Matthies. Galerkin-collocation approxi- mation in time for the wave equation and its post-processing. ESAIM Math. Model. Numer. Anal., 54(6):2099–2123, 2020

    work page 2020

  4. [4]

    A. K. Aziz and P. Monk. Continuous finite elements in space and time for the heat equation. Math. Comp. , 52(186):255–274, 1989

    work page 1989

  5. [5]

    Bales and I

    L. Bales and I. Lasiecka. Continuous finite elements in space and time for the nonhomogeneous wave equation. Comput. Math. Appl. , 27(3):91–102, 1994

    work page 1994

  6. [6]

    Bignardi and A

    P. Bignardi and A. Moiola. A space-time continuous and coercive formulation for the wave equation. arXiv:2312.07268, 2023

    arXiv 2023

  7. [7]

    Demkowicz, J

    L. Demkowicz, J. Gopalakrishnan, S. Nagaraj, and P. Sep´ ulveda. A spacetime DPG method for the Schr¨ odinger equation. SIAM J. Numer. Anal. , 55(4):1740–1759, 2017

    work page 2017

  8. [8]

    Z. Dong, L. Mascotto, and Z. Wang. A priori and a posteriori error estimates of a DG-CG method for the wave equation in second order formulation.arXiv:2411.03264, 2024

    arXiv 2024

Show all 30 references
  1. [9]

    Ern and J.-L

    A. Ern and J.-L. Guermond. Finite Elements I , volume 72 of Texts in Applied Mathematics. Springer Cham, 2021

    2021

  2. [10]

    L. C. Evans. Partial differential equations , volume 19 of Graduate Studies in Mathematics. American Mathematical Society, Providence, RI, second edition, 2010

    2010

  3. [11]

    M. Ferrari. XT-Waves-Exp-First-Order. https://github.com/MatteoFerrari11/ XT-Waves-Exp-First-Order.git , 2025. 23

    2025

  4. [12]

    Ferrari and S

    M. Ferrari and S. Fraschini. Stability of conforming space–time isogeometric methods for the wave equation. Math. Comp. , 2025. In press, doi: 10.1090/mcom/4062

    2025 doi

  5. [13]

    Ferrari, S

    M. Ferrari, S. Fraschini, G. Loli, and I. Perugia. Unconditionally stable space- time isogeometric discretization for the wave equation in Hamiltonian formulation. arXiv:2411.00650, 2024

    2024 arXiv

  6. [14]

    Ferrari, R

    M. Ferrari, R. L¨ oscher, and M. Zank. Stability of a modified Hilbert transform for second-order initial boundary value problems associated with space–time discretiza- tions of the wave equation. in preparation, 2025

    2025

  7. [15]

    Ferrari and I

    M. Ferrari and I. Perugia. Space-time discretization of the wave equation in a second-order-in-time formulation: a conforming, unconditionally stable method. arXiv:2503.11166, 2025

    2025 arXiv

  8. [16]

    Fraschini, G

    S. Fraschini, G. Loli, A. Moiola, and G. Sangalli. An unconditionally stable space–time isogeometric method for the acoustic wave equation. Comput. Math. Appl., 169:205–222, 2024

    2024

  9. [17]

    D. A. French. A space-time finite element method for the wave equation. Comput. Methods Appl. Mech. Engrg. , 107(1-2):145–157, 1993

    1993

  10. [18]

    D. A. French and T. E. Peterson. A continuous space-time finite element method for the wave equation. Math. Comp. , 65(214):491–506, 1996

    1996

  11. [19]

    F¨ uhrer, R

    T. F¨ uhrer, R. Gonz´ alez, and M. Karkulik. Well-posedness of first-order acoustic wave equations and space-time finite element approximation. IMA J. Numer. Anal. , page drae104, 03 2025

    2025

  12. [20]

    S. G´ omez. A variational approach to the analysis of the continuous space-time FEM for the wave equation. arXiv:2501.11494, 2025

    2025 arXiv

  13. [21]

    Henning, D

    J. Henning, D. Palitta, V. Simoncini, and K. Urban. An ultraweak space-time variational formulation for the wave equation: analysis and efficient numerical solution. ESAIM Math. Model. Numer. Anal. , 56(4):1173–1198, 2022

    2022

  14. [22]

    K¨ othe, R

    C. K¨ othe, R. L¨ oscher, and O. Steinbach. Adaptive least-squares space-time finite element methods. arXiv:2309.14300, 2023

    2023 arXiv

  15. [23]

    Lions and E

    J.-L. Lions and E. Magenes. Non-homogeneous boundary value problems and ap- plications. Vol. I , volume Band 181 of Die Grundlehren der mathematischen Wis- senschaften. Springer-Verlag, New York-Heidelberg, 1972. Translated from the French by P. Kenneth

    1972

  16. [24]

    L¨ oscher, O

    R. L¨ oscher, O. Steinbach, and M. Zank. Numerical results for an unconditionally stable space–time finite element method for the wave equation. In Domain De- composition Methods in Science and Engineering XXVI , pages 625–632. Springer, 2023

    2023

  17. [25]

    Lyche, C

    T. Lyche, C. Manni, and H. Speleers.Foundations of Spline Theory: B-Splines, Spline Approximation, and Hierarchical Refinement , pages 1–76. Springer International Publishing, Cham, 2018. 24

    2018

  18. [26]

    Steinbach

    O. Steinbach. Space-time finite element methods for parabolic problems. Comput. Methods Appl. Math. , 15(4):551–566, 2015

    2015

  19. [27]

    Steinbach and M

    O. Steinbach and M. Zank. A stabilized space–time finite element method for the wave equation, volume 128 of Lect. Notes Comput. Sci. Eng. Springer International Publishing, Cham, 2019

    2019

  20. [28]

    Steinbach and M

    O. Steinbach and M. Zank. Coercive space–time finite element methods for initial boundary value problems. Electron. Trans. Numer. Anal., 52:154–194, 2020

    2020

  21. [29]

    N. J. Walkington. Combined DG-CG time stepping for wave equations. SIAM J. Numer. Anal., 52(3):1398–1417, 2014

    2014

  22. [30]

    M. Zank. Higher-Order Space–Time Continuous Galerkin Methods for the Wave Equation. In 14th WCCM-ECCOMAS Congress 2020 , volume 700, 2021. 25

    2020

Pith tools

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