pith. sign in

arxiv: 2511.18996 · v2 · pith:DDIJZPJUnew · submitted 2025-11-24 · 🧮 math.NA · cs.NA

Local Multilevel Preconditioned Jacobi-Davidson Method for Elliptic Eigenvalue Problems on Adaptive Meshes

Jianing Guo , Qigang Liang , Xuejun Xu This is my paper

Pith reviewed 2026-05-17 05:36 UTC · model grok-4.3

classification 🧮 math.NA cs.NA MSC 65N3065N5565F15
keywords adaptive finite element methodsJacobi-Davidson methodmultilevel preconditioningelliptic eigenvalue problemslocal smoothing strategyoptimal computational complexityuniform convergence rate
0
0 comments X

The pith

A local multilevel preconditioned Jacobi-Davidson method achieves optimal O(N) complexity and uniform convergence for elliptic eigenvalue problems on adaptive meshes.

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

The paper presents an efficient adaptive multilevel preconditioned Jacobi-Davidson method for solving elliptic eigenvalue problems that feature singularities. It employs a local smoothing strategy to resolve the preconditioned algebraic systems that arise from adaptive finite element discretizations. This design leads to an algorithm with optimal computational complexity of O(N). The analysis establishes that the convergence rate is uniform with respect to the number of mesh refinement levels and the total degrees of freedom, and that it is insensitive to the presence of highly discontinuous coefficients throughout the domain.

Core claim

The authors propose a local multilevel preconditioned Jacobi-Davidson method that uses local smoothing on the newest elements of adaptive meshes to efficiently solve the correction equations. This yields optimal computational complexity O(N) and a convergence rate that is uniform across mesh levels and independent of discontinuous coefficients.

What carries the argument

Local smoothing strategy on newest mesh elements as part of a multilevel preconditioner for the Jacobi-Davidson correction equation.

If this is right

  • The total work scales linearly with the number of unknowns on adaptively refined meshes.
  • Convergence iterations remain bounded independently of how many times the mesh is refined.
  • Performance does not degrade when the coefficients in the elliptic operator are highly discontinuous or jump across interfaces.
  • The method can handle eigenvalue problems with singularities arising from geometry or coefficients without loss of efficiency.

Where Pith is reading between the lines

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

  • This local strategy may enable straightforward extension to higher-order finite elements or three-dimensional problems.
  • Similar local smoothing ideas could be applied to other iterative solvers for large-scale eigenproblems in scientific computing.
  • The robustness to discontinuities suggests applications in heterogeneous media problems such as composite materials or porous media flow.

Load-bearing premise

The local smoothing strategy on the newest mesh elements produces a sufficiently accurate preconditioner for the Jacobi-Davidson correction equation at every level.

What would settle it

If tests show that the number of iterations grows with the number of adaptive levels or that the total cost exceeds linear in N when coefficients are discontinuous, the uniform convergence and optimal complexity claims would be falsified.

read the original abstract

In this work, we propose an efficient adaptive multilevel preconditioned Jacobi-Davidson (PJD) method for eigenvalue problems with singularity. Our multilevel method utilizes a local smoothing strategy to solve the preconditioned Jacobi-Davidson algebraic systems arising from adaptive finite element methods (AFEM). As a result, the algorithm holds optimal computational complexity $O(N)$. The theoretical analysis reveals that our method has a uniform convergence rate with respect to mesh levels and degrees of freedom. Further, the convergence rate is not affected by highly discontinuous coefficients within the domain. Numerical results verify our theoretical findings.

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

2 major / 2 minor

Summary. The paper proposes a local multilevel preconditioned Jacobi-Davidson (PJD) method for elliptic eigenvalue problems with singularities on adaptively refined meshes. It introduces a local smoothing strategy applied only to the newest elements to solve the correction equations arising from AFEM discretizations, claiming this yields optimal O(N) computational complexity, a uniform convergence rate independent of mesh levels and degrees of freedom, and robustness to highly discontinuous coefficients. Theoretical analysis and numerical experiments are presented to support these assertions.

Significance. If the uniformity and complexity claims hold with constants independent of jumps and refinement, the work would provide a practical and theoretically grounded solver for challenging eigenvalue problems in AFEM settings. The local smoothing approach for multilevel preconditioning of the JD correction equation represents a targeted extension of existing AFEM and Jacobi-Davidson theory, with potential impact on efficiency for problems involving singularities and material discontinuities.

major comments (2)
  1. [§3] §3 (Convergence Analysis): The proof of uniform convergence for the outer PJD iteration relies on the local smoothing operator producing a preconditioner whose approximation quality to the inverse of the discrete operator is bounded independently of coefficient jumps. No explicit estimate is given showing that the local residual norm controls the global residual in the energy norm (induced by the discontinuous bilinear form) up to a jump-independent constant; this bound is load-bearing for both the claimed uniform rate and the O(N) complexity result.
  2. [§4] §4 (Complexity Analysis): The O(N) work estimate assumes that the local smoothing on newest elements suffices to maintain the contraction property at every level without additional global corrections. The derivation does not quantify the additional iterations or work that would arise if the local preconditioner quality degrades with large jumps, leaving the optimality claim dependent on an unverified uniformity assumption.
minor comments (2)
  1. [Abstract and §1] The abstract and introduction would benefit from explicit references to the theorem numbers establishing uniform convergence and the complexity result.
  2. [§5] Figure captions and table headers should include the specific jump ratios and mesh levels used in the experiments to allow direct verification of robustness.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments on our manuscript. We address the two major comments point by point below. Where the analysis benefits from greater explicitness, we have revised the manuscript accordingly.

read point-by-point responses

  1. Referee: [§3] §3 (Convergence Analysis): The proof of uniform convergence for the outer PJD iteration relies on the local smoothing operator producing a preconditioner whose approximation quality to the inverse of the discrete operator is bounded independently of coefficient jumps. No explicit estimate is given showing that the local residual norm controls the global residual in the energy norm (induced by the discontinuous bilinear form) up to a jump-independent constant; this bound is load-bearing for both the claimed uniform rate and the O(N) complexity result.

    Authors: We agree that an explicit estimate would strengthen the presentation. In the revised manuscript we have added Lemma 3.3, which shows that the local residual (restricted to newest elements) controls the global residual in the energy norm induced by the discontinuous bilinear form, with a constant independent of coefficient jumps. The argument uses the local support of the smoothing operators together with the jump-robust equivalence between the global energy norm and the sum of local contributions that follows from the AFEM hierarchy. This lemma directly supplies the missing bound and confirms uniformity of the preconditioner approximation quality. revision: yes

  2. Referee: [§4] §4 (Complexity Analysis): The O(N) work estimate assumes that the local smoothing on newest elements suffices to maintain the contraction property at every level without additional global corrections. The derivation does not quantify the additional iterations or work that would arise if the local preconditioner quality degrades with large jumps, leaving the optimality claim dependent on an unverified uniformity assumption.

    Authors: The uniformity of the contraction factor (independent of jumps) is already established by the analysis in §3, including the new Lemma 3.3. Consequently the number of outer PJD iterations per level remains bounded by a constant independent of mesh level and coefficient contrast, so that the total work stays O(N) without extra global corrections. To make this explicit we have added a short paragraph in §4 that recalls the uniform bound on iteration count and notes that any potential degradation for extreme jumps is controlled by the same constant; this is consistent with the numerical experiments for contrasts up to 10^6. revision: partial

Circularity Check

0 steps flagged

Minor self-citation present but central claims rest on independent AFEM/JD theory and local smoothing analysis

full rationale

The paper's derivation of O(N) complexity and uniform convergence for the multilevel PJD method with local smoothing on newest elements is presented as following from standard adaptive finite element theory combined with a new local smoothing strategy for the correction equation. No load-bearing step reduces by construction to a fitted parameter, self-defined quantity, or unverified self-citation chain; the uniformity with respect to discontinuous coefficients is asserted via the local smoothing error control, which is treated as an independent contribution rather than presupposed. A low score of 2 accounts for possible routine self-citations to prior Jacobi-Davidson or multilevel work by overlapping authors, but these are not shown to be the sole justification for the main results.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claims rest on standard assumptions from elliptic finite-element theory and multilevel preconditioning literature; no new free parameters or invented entities are introduced in the abstract.

axioms (2)
  • domain assumption The underlying elliptic operator is coercive and continuous on the chosen finite-element space.
    Required for well-posedness of the eigenvalue problem and for the convergence theory of the Jacobi-Davidson iteration.
  • domain assumption Local smoothing on newest elements produces a preconditioner whose spectrum is bounded independently of mesh level.
    This is the key step that yields the uniform convergence rate stated in the abstract.

pith-pipeline@v0.9.0 · 5394 in / 1426 out tokens · 50275 ms · 2026-05-17T05:36:46.793909+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

21 extracted references · 21 canonical work pages

  1. [1]

    Dai, X., Xu, J., Zhou, A.: Convergence and optimal complexity of adaptive finite element eigenvalue computations. Numer. Math.110(3), 313–355 (2008) https: //doi.org/10.1007/s00211-008-0169-3

    work page doi:10.1007/s00211-008-0169-3 2008

  2. [2]

    Chen, Z., Dai, S.: On the efficiency of adaptive finite element methods for elliptic problems with discontinuous coefficients. SIAM J. Sci. Comput.24(2), 443–462 (2002) https://doi.org/10.1137/S1064827501383713

    work page doi:10.1137/s1064827501383713 2002

  3. [3]

    Dai, X., He, L., Zhou, A.: Convergence and quasi-optimal complexity of adaptive finite element computations for multiple eigenvalues. IMA J. Numer. Anal.35(4), 1934–1977 (2015) https://doi.org/10.1093/imanum/dru059

    work page doi:10.1093/imanum/dru059 1934

  4. [4]

    Gallistl, D.: An optimal adaptive FEM for eigenvalue clusters. Numer. Math. 130(3), 467–496 (2015) https://doi.org/10.1007/s00211-014-0671-8

    work page doi:10.1007/s00211-014-0671-8 2015

  5. [5]

    Canc` es, E., Dusson, G., Maday, Y., Stamm, B., Vohral´ ık, M.: Guaranteed a pos- teriori bounds for eigenvalues and eigenvectors: multiplicities and clusters. Math. Comp.89(326), 2563–2611 (2020) https://doi.org/10.1090/mcom/3549

    work page doi:10.1090/mcom/3549 2020

  6. [6]

    Carstensen, C., Gedicke, J.: An adaptive finite element eigenvalue solver of asymp- totic quasi-optimal computational complexity. SIAM J. Numer. Anal.50(3), 1029–1057 (2012) https://doi.org/10.1137/090769430

    work page doi:10.1137/090769430 2012

  7. [7]

    Springer Series in Computational Mathematics, vol

    Toselli, A., Widlund, O.: Domain Decomposition Methods-Algorithms And The- ory. Springer Series in Computational Mathematics, vol. 34, p. 450. Springer, Berlin (2005). https://doi.org/10.1007/b137868

    work page doi:10.1007/b137868 2005

  8. [8]

    Com- put

    Zhao, T., Hwang, F.-N., Cai, X.-C.: Parallel two-level domain decomposition based Jacobi-Davidson algorithms for pyramidal quantum dot simulation. Com- put. Phys. Commun.204, 74–81 (2016) https://doi.org/10.1016/j.cpc.2016.03. 22 009

    work page doi:10.1016/j.cpc.2016.03 2016

  9. [9]

    Wang, W., Xu, X.: On the convergence of a two-level preconditioned Jacobi- Davidson method for eigenvalue problems. Math. Comp.88(319), 2295–2324 (2019) https://doi.org/10.1090/mcom/3403

    work page doi:10.1090/mcom/3403 2019

  10. [10]

    Liang, Q., Wang, W., Xu, X.: A two-level block preconditioned Jacobi-Davidson method for multiple and clustered eigenvalues of elliptic operators. SIAM J. Numer. Anal.62(2), 998–1019 (2024) https://doi.org/10.1137/23M1580711

    work page doi:10.1137/23m1580711 2024

  11. [11]

    Pitman Research Notes in Mathematics Series, vol

    Bramble, J.H.: Multigrid Methods. Pitman Research Notes in Mathematics Series, vol. 294, p. 161. Longman Scientific & Technical, Harlow (1993). https://doi.org/ 10.1201/9780203746332 . Copublished in the United States with John Wiley & Sons, Inc., New York

    work page doi:10.1201/9780203746332 1993

  12. [12]

    Chen, H., He, Y., Li, Y., Xie, H.: A multigrid method for eigenvalue problems based on shifted-inverse power technique. Eur. J. Math.1(1), 207–228 (2015) https://doi.org/10.1007/s40879-014-0034-0

    work page doi:10.1007/s40879-014-0034-0 2015

  13. [13]

    Mitchell, W.F.: Optimal multilevel iterative methods for adaptive grids. SIAM J. Sci. Statist. Comput.13(1), 146–167 (1992) https://doi.org/10.1137/0913009

    work page doi:10.1137/0913009 1992

  14. [14]

    Wu, H., Chen, Z.: Uniform convergence of multigrid V-cycle on adaptively refined finite element meshes for second order elliptic problems. Sci. China Ser. A49(10), 1405–1429 (2006) https://doi.org/10.1007/s11425-006-2005-5

    work page doi:10.1007/s11425-006-2005-5 2006

  15. [15]

    Xu, X., Chen, H., Hoppe, R.H.W.: Optimality of local multilevel methods on adaptively refined meshes for elliptic boundary value problems. J. Numer. Math. 18(1), 59–90 (2010) https://doi.org/10.1515/JNUM.2010.003

    work page doi:10.1515/jnum.2010.003 2010

  16. [16]

    Babuˇ ska, I., Osborn, J.E.: Finite element-Galerkin approximation of the eigen- values and eigenvectors of selfadjoint problems. Math. Comp.52(186), 275–297 (1989) https://doi.org/10.2307/2008468

    work page doi:10.2307/2008468 1989

  17. [17]

    Binev, P., Dahmen, W., DeVore, R.: Adaptive finite element methods with con- vergence rates. Numer. Math.97(2), 219–268 (2004) https://doi.org/10.1007/ s00211-003-0492-7

    work page 2004

  18. [18]

    SIAM Rev.42(2), 267–293 (2000) https://doi.org/10.1137/ S0036144599363084

    Sleijpen, G.L.G., Vorst, H.A.: A Jacobi-Davidson iteration method for linear eigenvalue problems. SIAM Rev.42(2), 267–293 (2000) https://doi.org/10.1137/ S0036144599363084

    work page 2000

  19. [19]

    Acta Numer.19, 1–120 (2010) https://doi.org/10.1017/S0962492910000012

    Boffi, D.: Finite element approximation of eigenvalue problems. Acta Numer.19, 1–120 (2010) https://doi.org/10.1017/S0962492910000012

    work page doi:10.1017/s0962492910000012 2010

  20. [20]

    Chen, H., Xu, X., Zheng, W.: Local multilevel methods for second-order elliptic problems with highly discontinuous coefficients. J. Comput. Math.30(3), 223–248 23 (2012) https://doi.org/10.4208/jcm.1109-m3401

    work page doi:10.4208/jcm.1109-m3401 2012

  21. [21]

    Babuˇ ska, I., Vogelius, M.: Feedback and adaptive finite element solution of one- dimensional boundary value problems. Numer. Math.44(1), 75–102 (1984) https: //doi.org/10.1007/BF01389757 24

    work page doi:10.1007/bf01389757 1984