Pith. sign in

REVIEW 3 major objections 4 minor 54 references

Variational Principles for the Helmholtz equation: application to Finite Element and Neural Network approximations

T0 review · 3 major / 4 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read Minimizing a regularized physical energy recovers the unique Helmholtz solution.

desk verdict The coercivity estimates look careful, but the central variational principle is false: a missing boundary term in the first variation means the exact Helmholtz solution is not stationary for Eγ or Fγ. read the letter →

arxiv 2511.13217 v2 pith:22OWCK6H submitted 2025-11-17 math.NA cs.NA

G. Makrakis , C. Makridakis , D. Mitsoudis , M. Plexousakis , T. Pryer This is my paper · ORCID
classification math.NAcs.NA MSC 35J0565N3065N35
keywords HelmholtzequationimpedanceboundaryconditionvariationalprinciplestrongcoercivityRellichidentityMorawetzfiniteelementmethodphysics-informedneuralnetworks
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 asks whether the Helmholtz equation with absorbing (impedance) boundary conditions can be derived from an energy principle. Starting from the least-action principle for the wave equation, the authors obtain the usual time-harmonic energy; it is sign-indefinite, so it cannot be minimized directly. The central claim is that adding least-squares penalties for the interior residual and for the impedance boundary condition, with sufficiently large weights, produces a strongly coercive energy whose unique minimizer is exactly the weak solution of the Helmholtz problem. Coercivity is proved with Rellich–Morawetz identities that control the indefinite bulk terms and boundary contributions, and the resulting energy supports both conforming finite-element discretizations with wavenumber-independent error bounds and a plane-wave neural network formulation.

What carries the argument

The key identity is the Rellich–Morawetz identity for the Helmholtz operator, which relates an integral of (x·∇u)Lu plus bulk gradient and boundary terms to zero; a second low-order identity with multiplier M0u = −ikβu controls the boundary terms. The paper uses these identities to show that a weighted combination of the physical energy and the least-squares residual has positive coefficients on ∥∇u∥², k²∥u∥², ∥Lu∥², and boundary terms whenever the penalty parameters γ1, γ2 are large enough. The regularized energies themselves carry the argument: Eγ restricts the trial space to functions satisfying the impedance condition, while Fγ adds a boundary residual term so the condition is enforced w

What would settle it

Take the one-dimensional Helmholtz problem on an interval with a known solution, e.g. u = e^{ikx}, and a smooth test function w in V_BC that is not proportional to u. Evaluate the first variation of Eγ at u against w; if the surviving boundary contribution Re(∂n u ̄w)|_0^1 is nonzero, the exact solution is not a stationary point over V_BC, contradicting the theorem. A discrete version: solve the variational problem (35) with conforming H² elements for a manufactured solution and measure whether the boundary impedance residual ∥∂n u_h − iku_h∥_{L²(∂Ω)} tends to zero under mesh refinement.

Watch

Extended reading notes

Core claim

The paper constructs two regularized energies, Eγ on the space of functions that already satisfy the impedance condition and Fγ on the larger space V where that condition is imposed weakly by a boundary penalty. Theorems 2 and 3 assert that for γ above computable thresholds the quadratic parts of these energies are strongly coercive in the respective norms, and therefore the minimization problems are well posed and the unique minimizers coincide with the weak solution of the Helmholtz equation. The proof uses Rellich and Morawetz identities, which convert the indefinite kinetic-minus-potential term into positive bulk and boundary contributions once the residual penalty is present. The same v

Load-bearing premise

The load-bearing premise is that stationarity of the regularized energy over the impedance-constrained space forces both the interior Helmholtz equation and the impedance boundary condition; the proof tests only compactly supported variations and never evaluates the boundary Euler–Lagrange term, so the boundary part of that premise is assumed rather than established.

Editorial extensions

If this is right

  • For γ above explicit thresholds, Eγ and Fγ are strictly convex and each has a unique minimizer, making the Helmholtz problem a well-posed minimization problem.
  • The coercivity constant is independent of k, removing the k-dependent inf-sup degradation that plagues standard Helmholtz finite element discretizations.
  • A conforming H² finite element discretization inherits coercivity and satisfies a quasi-optimal error bound with a constant independent of mesh size and wavenumber.
  • The same energy provides a neural-network loss that is an alternative to residual-based physics-informed formulations for high-wavenumber Helmholtz problems.
  • The regularization terms vanish on the solution manifold Lu = f, so at the solution the physical energy is recovered and the variational principle keeps its physical meaning.

Reading between the lines

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

  • Editorial inference: the 'physical energy + residual penalty + boundary penalty' template should transfer to other sign-indefinite time-harmonic boundary problems whenever a Rellich-type multiplier identity is available; the paper does not state this extension.
  • Editorial inference: because the coercivity and continuity constants are explicit and independent of k, the energy could serve as a certified loss for adaptive sampling or a posteriori error estimation in neural solvers; the paper only reports proof-of-concept experiments.
  • Editorial inference: the streamed least-squares initialization of the plane-wave network suggests a general warm-start strategy for variational neural solvers, separating the linear high-frequency part from the nonlinear correction; this is not developed in the paper.
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

3 major / 4 minor

Summary. The paper proposes variational principles for the Helmholtz equation with impedance boundary conditions. Starting from a time-averaged Lagrangian, it derives an indefinite physical energy EP, then augments it with least-squares residual terms, obtaining Eγ over the impedance-constrained space VBC and Fγ with a weak boundary penalty over V. The main theoretical claim is that, for sufficiently large penalty parameters, these augmented energies are strongly coercive and their unique minimisers coincide with the weak solution of the Helmholtz problem (1). Coercivity is proved via Rellich/Morawetz identities, and the framework is illustrated with H²-conforming finite element and plane-wave neural network discretisations.

Significance. If the central equivalence were correct, the paper would offer a physically motivated, coercive variational formulation for the impedance Helmholtz problem with potential value for finite element and neural network methods. The coercivity estimates in Section 4 are nontrivial and point to a useful set of energy identities. However, the central claim fails: the true solution is not a stationary point of Eγ or Fγ because a boundary term survives integration by parts. This is a load-bearing mathematical error, not a presentation issue. The unique minimisers of the proposed energies solve a different boundary-value problem, so the main theorems and the numerical methods built on them do not deliver what is claimed. The paper is therefore not suitable for publication in its current form.

major comments (3)
  1. [Section 3.2, Eq. (33)-(35) and Proposition 1] The stationarity condition (35) is not satisfied by the true solution u of (1). For u with Lu=f and ∂nu=iku, the first variation of Eγ at u is the same as that of the physical part, since the residual term vanishes. Using Green's identity, ⟨DEγ(u),v⟩ = Re{∫_Ω(∇u·∇v̄ - k²uv̄ - f v̄)} + boundary terms = Re{∫_∂Ω ∂nu v̄} = Re{∫_∂Ω iku v̄}. This does not vanish for all v∈VBC; choosing v=iu∈VBC gives Re{∫_∂Ω iku (-iū)} = k∫_∂Ω|u|² > 0. Thus u is not stationary, contradicting Theorem 2(ii)-(iii). The statement that the impedance condition is 'encoded in VBC' confuses the admissible test space with the boundary condition satisfied by the solution; the Euler-Lagrange derivation in Section 3.2 only tests with compactly supported functions and never checks the boundary contribution.
  2. [Section 3.2, Eq. (40)-(41) and Theorem 3] The same defect invalidates the weakly penalised energy Fγ. At the exact solution u, the boundary residual ∂nu-iku=0, so the γ2 term contributes nothing to the first variation; the physical part again leaves the boundary term Re{∫_∂Ω iku v̄}, which is nonzero for general v∈V. Hence the stationarity equation (41) is not equivalent to (1). In fact, for smooth solutions satisfying the interior equation, the boundary term would force a homogeneous Neumann-type condition rather than the impedance condition on the unconstrained space V. Consequently Theorem 3(ii)-(iii) is false, and the finite element and neural network minimisations in Section 5 target a different boundary-value problem.
  3. [Section 4, Theorems 14 and 15] The coercivity estimates (73) and (77) are statements about the quadratic parts Eγ(u)|_{f=0} and Fγ(u)|_{f=0}; at best they prove well-posedness of the variational problems (35) and (41). They do not establish that the unique solutions of those variational problems satisfy the impedance boundary condition. The paper's transition from coercivity to 'unique minimiser equals the solution of (1)' relies entirely on the false equivalence between the stationarity equations and problem (1). Since the true solution is not stationary for the proposed functionals, the coercivity results cannot rescue the central claim. The proofs of Theorem 2 and Theorem 3 therefore have a load-bearing gap at the point where Lax-Milgram is invoked.
minor comments (4)
  1. [Section 3.1-3.2, Eq. (25) and Eq. (33)] The real inner product is defined as ⟨φ,ψ⟩_R = Re∫_Ω φψ dx, but the first variation of |v|² in a complex Hilbert space should involve Re∫_Ω φ ψ̄ dx. Equations (25), (33) and (34) use unconjugated products; this inconsistency should be fixed, although the missing-boundary-term criticism above is independent of the conjugation convention.
  2. [Proposition 7, Eq. (57)] The proof of the low-order Morawetz identity appears to identify M0u with +ikβu at one step while the definition (56) sets M0u=-ikβu. Please check the sign consistency in the derivation; the final identity may still be correct, but the proof as written is confusing.
  3. [Section 5.4 and Figure 2] The neural network experiments show snapshots of the field but no comparison against a reference solution or an exact solution. Since the variational problem solved is not the impedance Helmholtz problem, the numerical evidence does not support the claim that the method approximates (1).
  4. [Abstract and Section 3.2] The abstract states that the unique minimisers coincide with solutions of the Helmholtz equation, but Proposition 1 itself only proves the interior equation from compactly supported variations. The boundary condition is asserted without a stationarity calculation on the boundary. This gap should be acknowledged at the statement of Proposition 1.

Circularity Check

1 steps flagged · score 3.0 of 10

The impedance boundary condition is inserted into the trial space V_BC by definition; the claimed 'stationary point = solution' is therefore partly true by construction, and the paper omits the boundary Euler–Lagrange term that would verify it.

  1. self definitional [Section 3.2, Proposition 1 (VBC definition; reused in Theorems 2–3, Eqs. (30), (38))]
    "The homogeneous-impedance subspace is VBC := {v ∈ V : ∂nv−ikv = 0 in L2(∂Ω)} ... Furthermore, if the energy EP restricted on V_BC has a stationary point, then this is the solution of the Helmholtz problem (1) where the impedance condition is encoded in the space V_BC."

    The target impedance condition is not derived from first variation; it is imposed as the definition of V_BC. The first variation's boundary term Re∫∂Ω ∂nu v̄ is never zeroed by an Euler–Lagrange equation; for the exact solution u of (1), it equals Re∫∂Ω iku v̄, which is not zero for all v∈V_BC (e.g. v=iu gives k‖u‖²_{L²(∂Ω)}>0). Hence the boundary part of 'stationary point equals solution of (1)' is an assumption built into the admissible space, not a consequence of stationarity; Theorem 2/3 inherit this by minimizing over the same V_BC (or penalizing ∂nv−ikv, which vanishes at u).

full rationale

The coercivity proofs via Rellich/Morawetz identities and Lax–Milgram are self-contained and constitute independent mathematical content; there is no fitted parameter being renamed as a prediction and no load-bearing self-citation (Moiola & Spence [38] are external and used mainly for comparison). The circular element is localized: the boundary condition is encoded by definition into V_BC, and the residual penalty is constructed to vanish exactly on the solution manifold, so the asserted coincidence of minimizer and solution is partly endowed by construction. At the same time, the exact solution is not stationary because the boundary term Re∫∂Ω iku v̄ survives; this is a correctness gap adjacent to, but not identical with, circularity. Score 3 reflects a moderate, partial circularity rather than a fully construction-forced result.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central claim depends on the star-shaped domain assumption, a regularity assumption on the solution, and several hand-chosen constants in the coercivity proof. No new physical entities are introduced. The most consequential assumption is the unproven regularity V_BC for the exact solution, but even granting it, the boundary-term omission breaks the main theorem.

free parameters (3)
  • γ (γ1, γ2) = γ1=39.5L², γ2=5.7L (ν=2 square); γ1=26.5L², γ2=5L (ν=3 cube)
    Regularization weights chosen by hand to satisfy the coercivity thresholds. They are not fitted to data but are free parameters of the method.
  • α, β, ε1, ε2, ε3 = e.g., ν=2: α=1, ε1=1/2, ε2=L0/4, ε3=L0/(4L²), β=6.2L
    Auxiliary constants introduced ad hoc in the coercivity proof to make Young's inequality terms positive. They are not data-derived.
  • λ (finite element boundary penalty) = λ=50
    User-specified consistent penalty parameter in the numerical experiments; chosen by hand.
assumptions (5)
  • domain assumption Assumption 4.1: Ω is strictly star-shaped with respect to the origin, x·n ≥ L0 > 0.
    Used throughout §4 to control boundary terms in the Rellich/Morawetz estimates. The experiments on the unit square [0,1]² do not satisfy this as stated unless coordinates are shifted.
  • domain assumption The weak solution u of (1) belongs to V_BC, i.e., ∂nu−iku∈L²(∂Ω) and u∈H¹(∂Ω).
    Asserted after (22) without proof. For general Lipschitz domains this regularity is not automatic from f∈L².
  • standard math Density of smooth functions in V_BC for the validity of Rellich identities.
    The coercivity proof is carried out for smooth functions and extended, but the density argument is not given.
  • standard math Lax–Milgram theorem for complex Hilbert spaces.
    Invoked to pass from coercivity to existence/uniqueness of the variational problem.
  • domain assumption Limiting Amplitude Principle for the time-harmonic ansatz.
    Used in §2.1 to justify that the time-periodic forcing leads to the Helmholtz equation; physically plausible but not proven.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Variational Principles for the Helmholtz equation: application to Finite Element and Neural Network approximations." pith.science (2026). https://pith.science/paper/22OWCK6H

@misc{pith2026251113217,
  author       = {Pith},
  title        = {Pith review of: Variational Principles for the Helmholtz equation: application to Finite Element and Neural Network approximations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/22OWCK6H}},
  note         = {Machine review of arXiv:2511.13217}
}
read the original abstract

In this paper, we investigate whether Variational Principles can be associated with the Helmholtz equation subject to impedance (absorbing) boundary conditions. This model has been extensively studied in the literature from both mathematical and computational perspectives. It is classical with wide applications, yet accurate approximation at high wavenumbers remains challenging. We address the question of whether there exist energy functionals with a clear physical interpretation whose stationary points, the zeros of their first variation, correspond to solutions of the Helmholtz problem. Starting from Hamilton's principle for the wave equation, we derive time-harmonic energies. The resulting functionals are generally indefinite. As a next step, we construct strongly coercive augmentations of these indefinite functionals that preserve their physical interpretation. Finally, we show how these variational principles lead to practical numerical methods based on finite element spaces and neural network architectures.

Figures

Figures reproduced from arXiv: 2511.13217 by the authors.

Figure 1
Figure 1. Real part of the H2 -conforming Argyris finite element solution on Ω = [0, 1]2 with h = 2−7 for wave numbers k ∈ {10, 50, 100, 200}. propagation directions and let {κr} R−1 r=0 be ring wavenumbers with κ0 = k and κr ≈ k for r ≥ 1. For x ∈ Ω define Φ cos r,p (x) = cos(κr ˆdp · x), Φ sin r,p(x) = sin(κr ˆdp · x), and stack them into the feature vector Φ(x) ∈ R D with D = 2P R. The model used in the experiments has two… view at source ↗
Figure 2
Figure 2. Plane-wave neural network approximation on [PITH_FULL_IMAGE:figures/full_fig_p023_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

54 extracted references · 1 canonical work pages

  1. [1]

    Coupled mode and finite element approximations of underwater sound propagation prob- lems in general stratified environments

    G. A. Athanassoulis, K. Belibassakis, D. A. Mitsoudis, N. Kampanis, and V . A. Dougalis. “Coupled mode and finite element approximations of underwater sound propagation prob- lems in general stratified environments”.Journal of Computational Acoustics16.1 (2008), pp. 83–116

    2008

  2. [2]

    A two point boundary value problem with a rapidly oscillating solution

    A. K. Aziz, R. B. Kellogg, and A. B. Stephens. “A two point boundary value problem with a rapidly oscillating solution”.Numer. Math.53.1-2 (1988), pp. 107–121

    1988

  3. [3]

    Is the pollution effect of the FEM avoidable for the Helmholtz equation considering high wave numbers?

    I. M. Babu ˇska and S. A. Sauter. “Is the pollution effect of the FEM avoidable for the Helmholtz equation considering high wave numbers?”SIAM Review42.3 (2012), pp. 451–484

    2012

  4. [4]

    On accuracy conditions for the numerical compu- tation of waves

    A. Bayliss, C. Goldstein, and E. Turkel. “On accuracy conditions for the numerical compu- tation of waves”.J. Comput. Phys.59.3 (1985), pp. 396–404

    1985

  5. [5]

    A space–time continuous and coercive formulation for the wave equation

    P. Bignardi and A. Moiola. “A space–time continuous and coercive formulation for the wave equation”.Numer. Math.157.4 (2025), pp. 1211–1258

    2025

  6. [6]

    Wave-number-explicit bounds in time-harmonic scat- tering

    S. N. Chandler-Wilde and P. Monk. “Wave-number-explicit bounds in time-harmonic scat- tering”.SIAM Journal on Mathematical Analysis39.5 (2008), pp. 1428–1455

    2008

  7. [7]

    Cossetti and D

    L. Cossetti and D. Krejcirik.The virial theorem and the method of multipliers in spectral theory. 2024. arXiv:2407.12379 [math.SP]

    arXiv 2024

  8. [8]

    An efficient neural network method with plane wave activa- tion functions for solving Helmholtz equation

    T. Cui, Z. Wang, and X. Xiang. “An efficient neural network method with plane wave activa- tion functions for solving Helmholtz equation”.Computers & Mathematics with Applications 111 (2022), pp. 34–49

    2022

Show all 54 references
  1. [9]

    Sharp regularity coefficient estimates for complex-valued acous- tic and elastic Helmholtz equations

    P. Cummings and X. Feng. “Sharp regularity coefficient estimates for complex-valued acous- tic and elastic Helmholtz equations”.Mathematical Models and Methods in Applied Sciences 16.01 (2006), pp. 139–160

    2006

  2. [10]

    Evolution equations on non-flat waveguides

    P. D’Ancona and R. Racke. “Evolution equations on non-flat waveguides”.Arch. Ration. Mech. Anal.206.1 (2012), pp. 81–110

    2012

  3. [11]

    Dautray and J.-L

    R. Dautray and J.-L. Lions.Mathematical analysis and numerical methods for science and technology. Vol. 2. Functional and variational methods, With the collaboration of Michel Artola, Marc Authier, Philippe B´enilan, Michel Cessenat, Jean Michel Combes, H´el`ene Lan- 24 chon, ...

    1988

  4. [12]

    Frequency domain treatment of one-dimensional scalar waves

    J. Douglas Jr., J. E. Santos, D. Sheen, and L. S. Bennethum. “Frequency domain treatment of one-dimensional scalar waves”.Mathematical Models and Methods in Applied Sciences3.2 (1993), pp. 171–194

    1993

  5. [13]

    L. C. Evans.Partial Differential Equations. Graduate Studies in Mathematics 19. American Mathematical Society, Providence, RI, 2010

    2010

  6. [14]

    hp-discontinuous Galerkin methods for the Helmholtz equation with large wave number

    X. Feng and H. Wu. “hp-discontinuous Galerkin methods for the Helmholtz equation with large wave number”.Math. Comp.80.276 (2011), pp. 1997–2024

    2011

  7. [15]

    Variational methods for underwater acoustic problems

    G. J. Fix and S. P. Marin. “Variational methods for underwater acoustic problems”.J. Comput. Phys.28.2 (1978), pp. 253–270

    1978

  8. [16]

    I. M. Gelfand and S. V . Fomin.Calculus of variations. English. Prentice-Hall, Inc., Engle- wood Cliffs, NJ, 1963, pp. vii+232

    1963

  9. [17]

    Givoli.Numerical methods for problems in infinite domains

    D. Givoli.Numerical methods for problems in infinite domains. Elsevier, New York, 1992

    1992

  10. [18]

    A finite element method for solving Helmholtz type equations in waveguides and other unbounded domains

    C. I. Goldstein. “A finite element method for solving Helmholtz type equations in waveguides and other unbounded domains”.Mathematics of Computation39.160 (1982), pp. 309–324

    1982

  11. [19]

    Stability and finite element error analysis for the Helmholtz equa- tion with variable coefficients

    I. Graham and S. Sauter. “Stability and finite element error analysis for the Helmholtz equa- tion with variable coefficients”.Mathematics of Computation89.321 (2020), pp. 105–138

    2020

  12. [20]

    Grisvard.Elliptic problems in nonsmooth domains

    P. Grisvard.Elliptic problems in nonsmooth domains. V ol. 24. Monographs and Studies in Mathematics. Pitman (Advanced Publishing Program), Boston, MA, 1985, pp. xiv+410

    1985

  13. [21]

    On controllability methods for the Helmholtz equation

    M. J. Grote and J. H. Tang. “On controllability methods for the Helmholtz equation”.Journal of Computational and Applied Mathematics358 (2019), pp. 306–326

    2019

  14. [22]

    Wave decay for star-shaped obstacles inR 3: papers of Morawetz and Ralston revisited

    P. Hintz and M. Zworski. “Wave decay for star-shaped obstacles inR 3: papers of Morawetz and Ralston revisited”.Math. Proc. R. Ir. Acad.117A.2 (2017), pp. 47–62

    2017

  15. [23]

    A survey of Trefftz methods for the Helmholtz equation

    R. Hiptmair, A. Moiola, and I. Perugia. “A survey of Trefftz methods for the Helmholtz equation”.Building bridges: connections and challenges in modern approaches to numerical partial differential equations. V ol. 114. Lect. Notes Comput. Sci. Eng. Springer, [Cham], 2016, pp. 237–278

    2016

  16. [24]

    Ihlenburg.Finite Element Analysis of Acoustic Scattering

    F. Ihlenburg.Finite Element Analysis of Acoustic Scattering. V ol. 132. Applied Mathematical Sciences. Springer New York, NY, 1998. 25

    1998

  17. [25]

    Finite element solution of the Helmholtz equation with high wave number Part I: The h-version of the FEM

    F. Ihlenburg and I. M. Babu ˇska. “Finite element solution of the Helmholtz equation with high wave number Part I: The h-version of the FEM.”Computers & Mathematics with Applica- tions30.9 (1995), pp. 9–37

    1995

  18. [26]

    Exact non-reflecting boundary conditions

    J. B. Keller and D. Givoli. “Exact non-reflecting boundary conditions”.J. Comput. Phys.82.1 (1989), pp. 172–192

    1989

  19. [27]

    Lanczos.The Variational Principles of Mechanics

    C. Lanczos.The Variational Principles of Mechanics. V ol. No. 4. Mathematical Expositions. University of Toronto Press, Toronto, ON, 1949, pp. xxv+307

    1949

  20. [28]

    Nonlinear wave equations in infinite waveguides

    P. H. Lesky and R. Racke. “Nonlinear wave equations in infinite waveguides”.Comm. Partial Differential Equations28.7-8 (2003), pp. 1265–1301

    2003

  21. [29]

    Lions and E

    J.-L. Lions and E. Magenes.Non-homogeneous boundary value problems and applications. Vol. I. V ol. Band 181. Die Grundlehren der mathematischen Wissenschaften. Translated from the French by P. Kenneth. Springer-Verlag, New York-Heidelberg, 1972, pp. xvi+357

    1972

  22. [30]

    DeepXDE: a deep learning library for solving differential equations

    L. Lu, X. Meng, Z. Mao, and G. E. Karniadakis. “DeepXDE: a deep learning library for solving differential equations”.SIAM Rev.63.1 (2021), pp. 208–228

    2021

  23. [31]

    Analysis and finite element methods for a fluid- solid interaction problem in one dimension

    C. Makridakis, F. Ihlenburg, and I. Babu ˇska. “Analysis and finite element methods for a fluid- solid interaction problem in one dimension”.Mathematical Models and Methods in Applied Sciences6.8 (1996), pp. 1119–1141

    1996

  24. [32]

    Convergence analysis for finite element discretizations of the Helmholtz equation with Dirichlet-to-Neumann boundary conditions

    J. Melenk and S. Sauter. “Convergence analysis for finite element discretizations of the Helmholtz equation with Dirichlet-to-Neumann boundary conditions”.Mathematics of Com- putation79.272 (2010), pp. 1871–1914

    2010

  25. [33]

    On Generalized Finite Element Methods

    J. M. Melenk. “On Generalized Finite Element Methods”. PhD thesis. University of Mary- land, College Park, MD: University of Maryland, 1995

    1995

  26. [34]

    Nonlinear hyperbolic equations in infinite homo- geneous waveguides

    J. Metcalfe, C. D. Sogge, and A. Stewart. “Nonlinear hyperbolic equations in infinite homo- geneous waveguides”.Comm. Partial Differential Equations30.4-6 (2005), pp. 643–661

    2005

  27. [35]

    Near- and far-field boundary conditions for a finite element method for the Helmholtz equation in axisymmetric problems of underwater acoustics

    D. A. Mitsoudis. “Near- and far-field boundary conditions for a finite element method for the Helmholtz equation in axisymmetric problems of underwater acoustics”.Acta Acustica united with Acustica93.6 (2007), pp. 888–898

    2007

  28. [36]

    Helmholtz equation with artificial boundary conditions in a two-dimensional waveguide

    D. A. Mitsoudis, C. Makridakis, and M. Plexousakis. “Helmholtz equation with artificial boundary conditions in a two-dimensional waveguide”.SIAM Journal on Mathematical Anal- ysis44.6 (2012), pp. 4320–4344

    2012

  29. [37]

    A finite element method with nonlocal boundary condi- tions for the Helmholtz equation with complex wavenumber in stratified waveguides

    D. A. Mitsoudis and M. Plexousakis. “A finite element method with nonlocal boundary condi- tions for the Helmholtz equation with complex wavenumber in stratified waveguides”. Acta Acustica united with Acustica95.4 (2009), pp. 753–756. 26

    2009

  30. [38]

    Is the Helmholtz Equation Really Sign-Indefinite?

    A. Moiola and E. A. Spence. “Is the Helmholtz Equation Really Sign-Indefinite?”SIAM Review56.2 (2014), pp. 274–312. eprint:https://doi.org/10.1137/120901301

    2014 doi

  31. [39]

    B. L. Moiseiwitsch.Variational principles. Corrected reprint of the 1966 original. Dover Publications, Inc., Mineola, NY, 2004, pp. x+310

    1966

  32. [40]

    An inequality for the reduced wave operator and the jus- tification of geometrical optics

    C. S. Morawetz and D. Ludwig. “An inequality for the reduced wave operator and the jus- tification of geometrical optics”.Communications on Pure and Applied Mathematics21.2 (1968), pp. 187–203

    1968

  33. [41]

    On the modes of decay for the wave equation in the exterior of a reflecting body

    C. S. Morawetz. “On the modes of decay for the wave equation in the exterior of a reflecting body”.Proc. Roy. Irish Acad. Sect. A72 (1972), pp. 113–120

    1972

  34. [42]

    Morrey–Campanato estimates for Helmholtz equations

    B. Perthame and L. Vega. “Morrey–Campanato estimates for Helmholtz equations”.Journal of functional analysis164.2 (1999), pp. 340–355

    1999

  35. [43]

    Eliminating the pollution effect in Helmholtz problems by local subscale cor- rection

    D. Peterseim. “Eliminating the pollution effect in Helmholtz problems by local subscale cor- rection”.Math. Comp.86.305 (2017), pp. 1005–1036

    2017

  36. [44]

    Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations

    M. Raissi, P. Perdikaris, and G. E. Karniadakis. “Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations”.J. Comput. Phys.378 (2019), pp. 686–707

    2019

  37. [45]

    A refined finite element convergence theory for highly indefinite Helmholtz problems

    S. A. Sauter. “A refined finite element convergence theory for highly indefinite Helmholtz problems”.Computing78.139 (2006), pp. 101–115

    2006

  38. [46]

    Stability estimate for the Helmholtz equation with rapidly jumping coefficients

    S. Sauter and C. Torres. “Stability estimate for the Helmholtz equation with rapidly jumping coefficients”.Zeitschrift f ¨ur angewandte Mathematik und Physik69.139 (2018)

    2018

  39. [47]

    An Observation Concerning Ritz-Galerkin Methods with Indefinite Bilinear Forms

    A. H. Schatz. “An Observation Concerning Ritz-Galerkin Methods with Indefinite Bilinear Forms”.Mathematics of Computation28.128 (1974), pp. 959–962

    1974

  40. [48]

    Feasibility study on solving the Helmholtz equation in 3D with PINNs

    S. Schoder and F. Kraxberger. “Feasibility study on solving the Helmholtz equation in 3D with PINNs”.arXiv preprint arXiv:2403.06623(2024)

    2024 arXiv

  41. [49]

    A perfectly matched layer for the Helmholtz equation in a semi- infinite strip

    I. Singer and E. Turkel. “A perfectly matched layer for the Helmholtz equation in a semi- infinite strip”.J. Comput. Phys.201.2 (2004), pp. 439–465

    2004

  42. [50]

    A versatile framework to solve the Helmholtz equation using physics-informed neural networks

    C. Song, T. Alkhalifah, and U. B. Waheed. “A versatile framework to solve the Helmholtz equation using physics-informed neural networks”.Geophysical Journal International228.3 (Oct. 2021), pp. 1750–1762. eprint:https://academic.oup.com/gji/article- pdf/228/3/1750/41166773/ggab434.pdf

    2021

  43. [51]

    A new frequency- uniform coercive boundary integral equation for acoustic scattering

    E. A. Spence, S. N. Chandler-Wilde, I. G. Graham, and V . P. Smyshlyaev. “A new frequency- uniform coercive boundary integral equation for acoustic scattering”.Communications on 27 Pure and Applied Mathematics64.10 (2011), pp. 1384–1415. eprint:https://onlinel ibrary.wiley.com...

    2011 doi

  44. [52]

    Morawetz inequalities

    T. Tao. “Morawetz inequalities”.Notices Amer. Math. Soc.65:7 (2018), pp. 764–770

    2018

  45. [53]

    Tao.Nonlinear dispersive equations

    T. Tao.Nonlinear dispersive equations. V ol. 106. CBMS Regional Conference Series in Mathematics. Local and global analysis. Conference Board of the Mathematical Sciences, Washington, DC; by the American Mathematical Society, Providence, RI, 2006, pp. xvi+373

    2006

  46. [54]

    D. R. Yafaev.Scattering Theory: Some Old and New Problems. V ol. 1735. Lecture Notes in Mathematics. Springer-Verlag Berlin Heidelberg, 2000. 28

    2000

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.