REVIEW 4 major objections 2 minor
Adaptive Crouzeix-Raviart finite elements for the first eigenpair of $p$-Laplacian
T0 review · 4 major / 2 minor · reviewed 2026-08-06 · deepseek-v4-flash
Pith's one-line read This paper proves that the adaptive Crouzeix-Raviart finite element algorithm converges for the first Dirichlet eigenpair of the p-Laplacian, with vanishing error estimators and recovery of the eigenvalue and eigenfunction.
desk verdict A plausible and potentially useful convergence theorem for adaptive Crouzeix-Raviart methods on the p-Laplacian first eigenpair, but the abstract leaves the load-bearing compactness lemma unverifiable. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central object is the adaptive Crouzeix-Raviart finite element space, a nonconforming piecewise-linear element with edge-midpoint degrees of freedom, built over a sequence of locally refined meshes and paired with a residual-type error estimator. The load-bearing mechanism is a compactness property for these spaces on adaptively generated mesh sequences: it provides enough uniform control to extract convergent subsequences of discrete eigenfunctions and to transfer the vanishing of the estimators into convergence of the eigenvalues and of the eigenfunctions in the broken norm. The analysis also relies on the variational formulation of the first $p$-Laplacian eigenpair and on the fine-initial-mesh condition that keeps the relevant constants uniform.
What would settle it
Run the adaptive Crouzeix-Raviart algorithm on a domain with a known first eigenpair, such as the unit ball where the first $p$-Laplacian eigenfunction is radial, starting from increasingly coarse initial meshes. If for any admissible marking strategy the estimator sequence fails to vanish or the eigenvalue iterates converge to a value different from the known first eigenvalue, the convergence claim fails, or the fine-initial-mesh condition would need explicit quantification.
Extended reading notes
Core claim
On its own terms, the central claim is that the adaptive Crouzeix-Raviart finite element algorithm for the first Dirichlet eigenpair of the $p$-Laplacian, the nonlinear operator $\Delta_p u=\nabla\cdot(|\nabla u|^{p-2}\nabla u)$, is convergent. Starting from a fine enough initial mesh, the sequence of error estimators produced by the adaptive loop is shown to have vanishing limit, the sequence of approximate eigenvalues is shown to converge to the first eigenvalue, and the mesh-dependent broken-norm distance between the discrete eigenfunctions and the set of relevant continuous eigenfunctions is shown to tend to zero. The proof's novelty is a compactness property for Crouzeix-Raviart finite element spaces over sequences of adaptively generated meshes, which supplies the control needed to pass from estimator decay to eigenpair convergence despite the nonconformity of the elements.
Load-bearing premise
The proof needs a guarantee that every admissible adaptively refined mesh sequence keeps the discrete eigenfunctions under enough control to extract a convergent subsequence with a well-behaved limit, and the initial mesh must be fine enough for that guarantee to hold; if any refinement pattern escapes this control, the convergence argument breaks.
Editorial extensions
If this is right
- The adaptive algorithm's estimated error tends to zero, so the marking and refinement strategy cannot stall on an admissible mesh sequence.
- The computed first eigenvalues converge to the exact first eigenvalue of the $p$-Laplacian, not merely to a spurious limiting value.
- The discrete first eigenfunctions converge in the mesh-dependent broken norm to the set of relevant continuous eigenfunctions, giving a concrete sense in which the eigenfunction is recovered.
- The Crouzeix-Raviart nonconforming element is a viable adaptive scheme for nonlinear eigenvalue problems, not only for linear source or eigenvalue problems.
Reading between the lines
- Editorial inference: the same compactness-based argument may extend to other nonconforming elements, such as rotated $Q_1$ elements, or to higher eigenpairs of the $p$-Laplacian, because the proof appears tied to the compactness property rather than to the specific edge-midpoint structure of Crouzeix-Raviart elements.
- Editorial inference: the unquantified fine-initial-mesh condition could be made explicit by tracking the constants in the compactness proof, giving users a concrete threshold for how fine the starting mesh must be on a given domain.
- Editorial inference: if the estimator is also reliable and efficient, the vanishing-estimator result is a natural first step toward proving optimal convergence rates for the adaptive loop, although rates are not claimed in this paper.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes and analyzes an adaptive Crouzeix-Raviart finite element method for computing the first Dirichlet eigenpair of the p-Laplacian. The abstract claims (i) the sequence of error estimators produced by the adaptive algorithm has a vanishing limit, (ii) starting from a fine initial mesh, the approximate eigenvalues converge to the first eigenvalue, and (iii) the distance in a mesh-dependent broken norm between discrete eigenfunctions and a set of 'relevant continuous eigenfunctions' tends to zero. The analysis is said to hinge on a new compactness property for Crouzeix-Raviart spaces over adaptively generated meshes. Numerical experiments are presented as illustrations of the algorithm's advantage.
Significance. If the convergence results are correct, the paper would make a significant contribution to adaptive nonconforming finite element methods for nonlinear eigenvalue problems, specifically for the p-Laplacian where the nonlinearity and degeneracy complicate both approximation and a posteriori error control. The compactness property for Crouzeix-Raviart spaces over adaptively refined meshes is described as a key theoretical novelty; establishing such a property is genuinely nontrivial and, if proven, would be a valuable addition to the literature. The paper does not appear to use fitted parameters or circular benchmarks, and the convergence claims target the true eigenproblem. However, the abstract alone does not permit verification of the proofs or the numerical evidence: no error tables, baseline comparisons, or precise statements of the compactness lemma are provided.
major comments (4)
- [Abstract] The phrase 'starting from a fine initial mesh' is load-bearing for the eigenvalue convergence claim, yet no quantification is given. The theorem is stated conditional on this initialization, but the threshold for 'fine' (e.g., an upper bound on the initial mesh size h0 in terms of the domain, the exponent p, and the spectral gap) is absent. Without such a quantification, the statement is not a convergence result for the adaptive iteration from arbitrary initial meshes, but a statement about a family of initializations whose properties are left unspecified.
- [Abstract] The compactness property for Crouzeix-Raviart spaces over adaptively generated meshes is the stated hinge of the analysis, but its precise hypotheses are not enumerated. For nonconforming methods, discrete compactness is delicate because the broken gradient does not control interelement boundary jumps; establishing strong convergence in L^p from a boundedness-in-broken-norm argument typically requires additional geometric mesh constraints such as uniform shape regularity, a bounded number of element neighbors, or a fixed refinement ratio. The abstract does not state whether the compactness lemma holds for every admissible adaptive refinement pattern or only under such extra conditions, which is essential for assessing the validity of the convergence proof.
- [Abstract] The convergence statement 'the distance in a mesh-dependent broken norm between discrete eigenfunctions and the set composed of relevant continuous eigenfunctions also tends to zero' refers to a set of 'relevant continuous eigenfunctions' that is never defined in the abstract. Without a specification of this set (whether it is the full eigenspace for the first eigenvalue, a normalized subset, or some other collection), the statement cannot be checked or falsified, and it is also unclear how the mesh-dependent broken norm is defined.
- [Abstract] The numerical experiments are described only as illustrations of the advantage of the proposed algorithm, with no baseline comparisons, error tables, or convergence rates reported in the abstract. As a result, the numerical evidence cannot be independently assessed; if the full paper provides these details, this comment is a request to ensure that the experimental section includes quantitative comparisons against a standard (possibly uniform) method and reports the actual estimated convergence orders.
minor comments (2)
- [Abstract] The notation 'p-Laplacian problem' is ambiguous in an eigenvalue context; the abstract should specify that the Dirichlet eigenvalue problem is considered, namely find (lambda,u) with -div(|grad u|^{p-2} grad u) = lambda |u|^{p-2} u in Omega and u=0 on the boundary, for 1 < p < infinity.
- [Abstract] The term 'error estimators' is used without specifying the residual type or the norm in which the error is estimated; the abstract would benefit from a one-sentence definition of the estimator or a reference to the main text.
Circularity Check
No circularity found in the abstract-level derivation chain.
full rationale
This is an abstract-only review, so the derivation chain cannot be fully audited; however, nothing in the available text exhibits a circular step. The convergence claims target the true first Dirichlet eigenvalue and the set of continuous eigenfunctions, with the discrete objects compared in a mesh-dependent broken norm. No parameter is fitted to data and then renamed a prediction. No self-citation is invoked as load-bearing; the compactness property for Crouzeix-Raviart spaces over adaptively generated meshes is asserted as a proved result, not assumed as an input equivalent to the conclusion. The phrase 'starting from a fine initial mesh' is an initialization condition that is not quantified, but an unquantified condition is a verification gap or a potential correctness risk, not circularity. Under the hard rules, circularity requires quoting a specific reduction where a claimed output equals an input by construction; no such reduction is visible in the abstract. The honest non-finding is therefore a score of 0.
Assumptions & free parameters
assumptions (2)
- domain assumption The computational domain is a bounded Lipschitz domain and the p-Laplacian Dirichlet eigenproblem has a well-defined first eigenpair.
- domain assumption Admissible adaptive mesh refinements preserve shape regularity and support the Crouzeix-Raviart interpolation and compactness estimates used in the proof.
Cite this review
Pith. "Pith review of Adaptive Crouzeix-Raviart finite elements for the first eigenpair of $p$-Laplacian." pith.science (2026). https://pith.science/paper/3VF4MLZZ
@misc{pith2026250802077,
author = {Pith},
title = {Pith review of: Adaptive Crouzeix-Raviart finite elements for the first eigenpair of $p$-Laplacian},
year = {2026},
howpublished = {\url{https://pith.science/paper/3VF4MLZZ}},
note = {Machine review of arXiv:2508.02077}
}
abstract
In this paper, we propose and analyze an adaptive Crouzeix-Raviart finite element method for computing the first Dirichlet eigenpair of the $p$-Laplacian problem. We prove that the sequence of error estimators produced by the adaptive algorithm has a vanishing limit and that, starting from a fine initial mesh, the relevant sequence of approximate eigenvalues converges to the first eigenvalue and the distance in a mesh-dependent broken norm between discrete eigenfunctions and the set composed of relevant continuous eigenfunctions also tends to zero. The analysis hinges on establishing a compactness property for Crouzeix-Raviart finite elements over a sequence of adaptively generated meshes, which represents key theoretical challenges and novelties. We present numerical results to illustrate the advantage of the proposed algorithm.
Reviewed August 6, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.