pith. sign in

arxiv: 2403.18378 · v5 · submitted 2024-03-27 · 🧮 math.NA · cs.NA

Can Symmetric Positive Definite (SPD) coarse spaces perform well for indefinite Helmholtz problems?

Victorita Dolean , Mark Fry , Matthias Langer This is my paper

Pith reviewed 2026-05-24 03:44 UTC · model grok-4.3

classification 🧮 math.NA cs.NA
keywords Helmholtz equationdomain decompositionGenEOcoarse spaceSPDGMRESadditive Schwarzpreconditioner
0
0 comments X

The pith

Adapted Δ_k-GenEO coarse spaces sharpen GMRES convergence conditions for heterogeneous Helmholtz problems using SPD eigenvalue problems.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces the Δ_k-GenEO coarse space for two-level additive Schwarz methods applied to the indefinite Helmholtz equation with heterogeneous coefficients. It provides an analysis showing that this adaptation of the Δ-GenEO space, based on local symmetric positive definite eigenvalue problems, leads to improved k-explicit bounds for the convergence of GMRES. The sharpened conditions relax previous requirements on subdomain size and the number of eigenvectors retained. This helps explain why such SPD-based coarse spaces perform well in practice for wave problems even though the operator is indefinite. Readers would be interested because it bridges the gap between theoretical predictions and observed numerical behavior in domain decomposition for challenging wave propagation simulations.

Core claim

Within the two-level additive Schwarz framework for heterogeneous Helmholtz problems, the Δ_k-GenEO coarse space constructed from SPD local eigenvalue problems sharpens the k-explicit conditions for GMRES convergence. This reduces the restrictions on the subdomain size and the eigenvalue threshold, while numerical experiments demonstrate scalability, robustness to heterogeneity at low to moderate frequencies, and a milder growth of the coarse space dimension than predicted by earlier conservative estimates.

What carries the argument

The Δ_k-GenEO coarse space, an adaptation of the Δ-GenEO construction that uses symmetric positive definite eigenvalue problems to select modes relevant to the indefinite Helmholtz operator.

If this is right

  • The number of GMRES iterations is bounded under less restrictive conditions on the subdomain diameter relative to the wavelength.
  • The coarse space dimension increases more slowly with frequency than previously estimated.
  • Robustness with respect to variations in material properties holds for frequencies up to moderate values of k.
  • Scalability with respect to the number of subdomains is achieved in the tested regimes.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Similar adaptations could be tested on other indefinite operators such as those arising in electromagnetics.
  • The observed limitations at high frequencies point to potential benefits from combining with frequency-dependent coarse space strategies.
  • If the SPD problems capture the modes effectively, this may reduce the need for specialized indefinite eigenvalue solvers in coarse space construction.

Load-bearing premise

The SPD local eigenvalue problems capture the essential oscillatory behavior of the indefinite Helmholtz operator sufficiently well without explicit accounting for sign changes or heterogeneity effects.

What would settle it

Observing that the GMRES iteration count exceeds the sharpened theoretical bound for a heterogeneous Helmholtz problem with moderate k and subdomain size satisfying the new condition would falsify the sharpening claim.

Figures

Figures reproduced from arXiv: 2403.18378 by Mark Fry, Matthias Langer, Victorita Dolean.

Figure 1
Figure 1. Figure 1: Influence of the coarse space size (left) and threshold choice (right) on the iteration count for the homogeneous media test case with [PITH_FULL_IMAGE:figures/full_fig_p020_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Influence of the subdomain diameter on the iteration count (left) and coarse space size (right) for the homogeneous media test case [PITH_FULL_IMAGE:figures/full_fig_p021_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Influence of the coarse space size (left) and threshold choice (right) on the iteration count for the heterogeneous media test case with [PITH_FULL_IMAGE:figures/full_fig_p023_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Influence of the subdomain diameter on the iteration count (left) and coarse space size (right) for the heterogeneous media test case [PITH_FULL_IMAGE:figures/full_fig_p024_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: The heterogeneous function a(x) within the alternating layers. The shading gives the value of a(x) with the darkest shade being a(x) = amax, where amax > 1 is a parameter, and the white taking the value amin = 1. 5.3. Results In Tables 2 and 3, we display the results comparing the one-level AS and the two-level AS with the ∆-GenEO from [12] and the newly proposed ∆k-GenEO. The indefiniteness of the problem… view at source ↗
Figure 6
Figure 6. Figure 6: Influence of the coarse space size (left) and threshold choice (right) on the iteration count for the homogeneous media test case with [PITH_FULL_IMAGE:figures/full_fig_p029_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Influence of the subdomain diameter on the iteration count (left) and coarse space size (right) for the homogeneous media test case [PITH_FULL_IMAGE:figures/full_fig_p030_7.png] view at source ↗
read the original abstract

Wave propagation problems governed by the Helmholtz equation remain among the most challenging in scientific computing, due to their indefinite nature. Domain decomposition methods with spectral coarse spaces have emerged as some of the most effective preconditioners, yet their theoretical guarantees often lag behind practical performance. In this work, we introduce and analyse the $\Delta_k$-GenEO coarse space within the two-level additive Schwarz preconditioners for heterogeneous Helmholtz problems. This is an adaptation of the $\Delta$-GenEO coarse space. Our results sharpen the $k$-explicit conditions for GMRES convergence, reducing the restrictions on the subdomain size and eigenvalue threshold. This narrows the long-standing gap between pessimistic theory and empirical evidence, and reveals why GenEO spaces based on SPD (symmetric positive definite) eigenvalue problems remain surprisingly effective despite their apparent limitations. Numerical experiments confirm the theory, demonstrating scalability, robustness to heterogeneity for low to moderate frequencies (while experiencing limitations in the high frequency cases), and significantly milder coarse-space growth than conservative estimates predict.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 2 minor

Summary. The paper introduces the Δ_k-GenEO coarse space as an adaptation of the Δ-GenEO construction for two-level additive Schwarz preconditioners applied to heterogeneous Helmholtz problems. It claims to derive sharpened k-explicit conditions for GMRES convergence that reduce the restrictions on subdomain size and eigenvalue threshold relative to prior bounds. The work argues this narrowing of the theory-practice gap explains the observed effectiveness of SPD-based GenEO spaces for indefinite operators, supported by numerical experiments showing scalability and robustness to heterogeneity at low to moderate frequencies (with noted degradation at high frequencies).

Significance. If the sharpened bounds hold and the explanation for SPD coarse-space effectiveness is valid, the result would meaningfully advance the analysis of spectral coarse spaces for indefinite wave-propagation problems. It could provide a concrete mechanism for why GenEO constructions based on SPD eigenproblems succeed in practice despite the indefiniteness of the target operator, offering both theoretical insight and practical guidance on coarse-space dimension growth.

major comments (1)
  1. The central claim that the adapted Δ_k-GenEO construction using local SPD eigenvalue problems produces a coarse space controlling the oscillatory modes of the indefinite Helmholtz operator (thereby yielding the sharpened k-explicit GMRES bounds and relaxed subdomain-size/eigenvalue-threshold restrictions) is load-bearing. The abstract itself reports performance degradation at high frequencies, indicating that the assumption may fail to hold uniformly without explicit corrections for sign-indefiniteness or strong heterogeneity; if so, the claimed reduction in restrictions would not follow.
minor comments (2)
  1. The abstract should explicitly state the frequency ranges (in terms of k or kh) corresponding to 'low to moderate' versus 'high frequency' regimes to allow readers to assess the scope of the reported robustness.
  2. Notation for the Δ_k-GenEO space and its relation to the original Δ-GenEO should be introduced with a clear definition and comparison table early in the manuscript.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the careful review and the opportunity to address this important point on the scope of our theoretical claims.

read point-by-point responses

  1. Referee: The central claim that the adapted Δ_k-GenEO construction using local SPD eigenvalue problems produces a coarse space controlling the oscillatory modes of the indefinite Helmholtz operator (thereby yielding the sharpened k-explicit GMRES bounds and relaxed subdomain-size/eigenvalue-threshold restrictions) is load-bearing. The abstract itself reports performance degradation at high frequencies, indicating that the assumption may fail to hold uniformly without explicit corrections for sign-indefiniteness or strong heterogeneity; if so, the claimed reduction in restrictions would not follow.

    Authors: We agree the claim is central and that the theory is not uniform. The sharpened k-explicit GMRES bounds hold under explicit assumptions relating k to subdomain diameter and the eigenvalue threshold; these assumptions are relaxed relative to earlier analyses but still restrict the result to regimes where the coarse space controls the relevant modes. The numerical degradation at high frequencies is consistent with the theory once those assumptions are violated. The reduction in restrictions is therefore comparative (narrowing the theory-practice gap for low-to-moderate frequencies) rather than a claim of uniformity over all k. We will add a clarifying sentence in the introduction and abstract to emphasize the conditional character of the bounds. revision: partial

Circularity Check

0 steps flagged

No circularity: sharpened bounds presented as independent analysis of adapted GenEO

full rationale

The provided abstract and context describe an adaptation of the Δ-GenEO construction to Δ_k-GenEO for the indefinite Helmholtz operator, followed by sharpened k-explicit GMRES convergence conditions that reduce subdomain-size and eigenvalue-threshold restrictions. No equations or self-citation chains are exhibited that would make the sharpened bounds equivalent to the input eigenvalue threshold or partitioning choices by construction. The paper positions the numerical experiments as independent confirmation of the theory, and the central claim (SPD eigenproblems remain effective) is framed as an explanatory result rather than a renaming or fitted-input prediction. Absent explicit load-bearing reductions in the derivation chain, the analysis is self-contained.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No specific free parameters, axioms, or invented entities are described in the abstract; the review therefore records empty lists.

pith-pipeline@v0.9.0 · 5704 in / 1222 out tokens · 36944 ms · 2026-05-24T03:44:11.917084+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

24 extracted references · 24 canonical work pages

  1. [1]

    S. C. Eisenstat, H. C. Elman, M. H. Schultz, Variational iterative methods for nonsymmetric systems of linear equations, SIAM J. Numer. Anal. 20 (2) (1983) 345–357.doi:10.1137/0720023

    work page doi:10.1137/0720023 1983

  2. [2]

    X.-C. Cai, M. Sarkis, A restricted additive Schwarz preconditioner for general sparse linear systems, SIAM J. Sci. Comput. 21 (2) (1999) 792–797.doi:10.1137/S106482759732678X

    work page doi:10.1137/s106482759732678x 1999

  3. [3]

    Toselli, O

    A. Toselli, O. Widlund, Domain Decomposition Methods – Algorithms and Theory, V ol. 34 of Springer Series in Computational Mathematics, Springer, 2005

    work page 2005

  4. [4]

    Graham, P

    I. Graham, P. Lechner, R. Scheichl, Domain decomposition for multiscale PDEs, Numer. Math. 106 (2007) 589–626.doi:10.1007/s00211-007-0074-1

    work page doi:10.1007/s00211-007-0074-1 2007

  5. [5]

    Spillane, V

    N. Spillane, V . Dolean, P. Hauret, F. Nataf, C. Pechstein, R. Scheichl, Abstract robust coarse spaces for systems of PDEs via generalized eigenproblems in the overlaps, Numer. Math. 126 (4) (2014) 741–770.doi:10.1007/ s00211-013-0576-y

    work page 2014

  6. [6]

    Conen, V

    L. Conen, V . Dolean, R. Krause, F. Nataf, A coarse space for heterogeneous Helmholtz problems based on the Dirichlet-to-Neumann operator, J. Comput. Appl. Math. 271 (2014) 83–99.doi:10.1016/j.cam.2014.03. 031

    work page doi:10.1016/j.cam.2014.03 2014

  7. [7]

    Nataf, H

    F. Nataf, H. Xiang, V . Dolean, N. Spillane, A coarse space construction based on local Dirichlet-to-Neumann maps, SIAM J. Sci. Comput. 33 (4) (2011) 1623–1642.doi:10.1137/100796376

    work page doi:10.1137/100796376 2011

  8. [8]

    Dolean, F

    V . Dolean, F. Nataf, R. Scheichl, N. Spillane, Analysis of a two-level Schwarz method with coarse spaces based on local Dirichlet-to-Neumann maps, Comput. Methods Appl. Math. 12 (4) (2012) 391–414.doi:10.2478/ cmam-2012-0027

    work page 2012

  9. [9]

    Bootland, V

    N. Bootland, V . Dolean, Can DtN and GenEO coarse spaces be sufficiently robust for heterogeneous Helmholtz problems?, Math. Comput. Appl. 27 (3) (2022).doi:10.3390/mca27030035

    work page doi:10.3390/mca27030035 2022

  10. [10]

    Bootland, V

    N. Bootland, V . Dolean, P. Jolivet, P.-H. Tournier, A comparison of coarse spaces for Helmholtz problems in the high frequency regime, Comput. Math. Appl. 98 (2021) 239–253.doi:10.1016/j.camwa.2021.07.011

    work page doi:10.1016/j.camwa.2021.07.011 2021

  11. [11]

    Dolean, M

    V . Dolean, M. Fry, M. Langer, E. Parolin, P.-H. Tournier, Achieving wavenumber robustness in domain de- composition for heterogeneous Helmholtz equation: an overview of spectral coarse spaces (2025).arXiv: 2509.02131. 31

    work page arXiv 2025

  12. [12]

    Bootland, V

    N. Bootland, V . Dolean, I. G. Graham, C. Ma, R. Scheichl, Overlapping Schwarz methods with GenEO coarse spaces for indefinite and nonself-adjoint problems, IMA J. Numer. Anal. 43 (4) (2022) 1899–1936.doi:10. 1093/imanum/drac036

    work page 2022

  13. [13]

    Operto, P

    S. Operto, P. Amestoy, H. Aghamiry, S. Beller, A. Buttari, L. Combe, V . Dolean, M. Gerest, G. Guo, P. Jolivet, et al., Is 3D frequency-domain FWI of full-azimuth/long-offset OBN data feasible? The Gorgon data FWI case study, The Leading Edge 42 (3) (2023) 173–183

    work page 2023

  14. [14]

    Tournier, I

    P.-H. Tournier, I. Aliferis, M. Bonazzoli, M. de Buhan, M. Darbas, V . Dolean, F. Hecht, P. Jolivet, I. E. Kanfoud, C. Migliaccio, F. Nataf, C. Pichot, S. Semenov, Microwave tomographic imaging of cerebrovascular accidents by using high-performance computing, Parallel Computing 85 (2019) 88–97.doi:10.1016/j.parco.2019. 02.004

    work page doi:10.1016/j.parco.2019 2019

  15. [15]

    Tournier, P

    P.-H. Tournier, P. Jolivet, V . Dolean, H. S. Aghamiry, S. Operto, S. Riffo, 3D finite-difference and finite-element frequency-domain wave simulation with multilevel optimized additive Schwarz domain-decomposition precon- ditioner: a tool for full-waveform inversion of sparse node data sets, GEOPHYSICS 87 (5) (2022) T381–T402. doi:10.1190/geo2021-0702.1

    work page doi:10.1190/geo2021-0702.1 2022

  16. [16]

    C. Ma, C. Alber, R. Scheichl, Two-level restricted additive Schwarz preconditioner based on multiscale spectral generalized FEM for heterogeneous Helmholtz problems (2024).arXiv:2409.06533

    work page arXiv 2024

  17. [17]

    Schatz, J

    A. Schatz, J. Wang, Some new error estimates for Ritz–Galerkin methods with minimal regularity assumptions, Math. Comp. 65 (213) (1996) 19–27

    work page 1996

  18. [18]

    Ihlenburg, Finite Element Analysis of Acoustic Scattering, V ol

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

    work page 1998

  19. [19]

    J. M. Melenk, S. Sauter, Wavenumber explicit convergence analysis for galerkin discretizations of the helmholtz equation, SIAM Journal on Numerical Analysis 49 (3) (2011) 1210–1243.arXiv:https://doi.org/10. 1137/090776202,doi:10.1137/090776202. URLhttps://doi.org/10.1137/090776202

    work page doi:10.1137/090776202 2011

  20. [20]

    Leoni, A First Course in Sobolev Spaces, Graduate Studies in Mathematics, American Mathematical Society, 2017

    G. Leoni, A First Course in Sobolev Spaces, Graduate Studies in Mathematics, American Mathematical Society, 2017

    work page 2017

  21. [21]

    Bastian, R

    P. Bastian, R. Scheichl, L. Seelinger, A. Strehlow, Multilevel spectral domain decomposition, SIAM J. Sci. Comput. 45 (3) (2023) S1–S26.doi:10.1137/21M1427231

    work page doi:10.1137/21m1427231 2023

  22. [22]

    Tournier, P

    P.-H. Tournier, P. Jolivet, F. Nataf, FFDDM: FreeFem domain decomposition method (2019). URLhttps://doc.freefem.org/documentation/ffddm/index.html

    work page 2019

  23. [23]

    Amestoy, I

    P. Amestoy, I. Duff, J.-Y . L’Excellent, J. Koster, A fully asynchronous multifrontal solver using distributed dynamic scheduling, SIAM J. Matrix Anal. Appl. 23 (1) (2001) 15–41. URLhttp://mumps.enseeiht.fr

    work page 2001

  24. [24]

    Hernandez, J

    V . Hernandez, J. E. Roman, V . Vidal, SLEPc: A scalable and flexible toolkit for the solution of eigenvalue problems, ACM Transactions on Mathematical Software 31 (3) (2005) 351–362. URLhttps://slepc.upv.es 32

    work page 2005