Pith. sign in

REVIEW 4 major objections 7 minor 10 references

Universal estimates for the density of states for aperiodic block subwavelength resonator systems

T0 review · 4 major / 7 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper proves that for one-dimensional random block subwavelength resonator arrays, the integrated density of states converges almost surely to a non-random continuous function, and that the fractal-like hybridisation-region spectrum…

desk verdict A useful, clearly written paper that proves an ergodic limit for the discrete capacitance model and introduces an effective meta-atom algorithm; the gap to the physical resonances is unproven but probably fixable. read the letter →

arxiv 2505.16677 v1 pith:5ANOISFP submitted 2025-05-22 math-ph cond-mat.dis-nncond-mat.mtrl-scimath.MP

Habib Ammari , Silvio Barandun , Bryn Davies , Erik Orvehed Hiltunen , Alexander Uhlmann This is my paper · ORCID
classification math-phcond-mat.dis-nncond-mat.mtrl-scimath.MP MSC 35J0535C2035P20
keywords densityofstatesblockdisorderedsystemssubwavelengthresonatorshybridisationregionsmetrictransitivityJacobioperatorsquasiperiodicsamplinghyperuniform
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 studies long chains of subwavelength acoustic resonators built by sampling independently from a small set of building blocks. It sets out to show that, once the chain is long, the distribution of resonant frequencies becomes a single non-random, continuous curve, independent of the particular random sample; numerically it finds that this limiting density of states splits into three regimes: zero in bandgaps, smooth bands where every block type passes, and a fractal-like set of peaks in hybridisation regions where some block types are gapped. It goes on to explain that the fractal-like peaks come from eigenmodes decaying exponentially through the gapped blocks, and uses that insight to build a meta-atom estimator that reconstructs the hybridisation-region spectrum in time linear in the number of blocks. A reader should care because it says a deterministic, computable spectral law governs nominally disordered acoustic metamaterials, and the exotic fractal-looking spectrum is a finite-size signature rather than the infinite limit.

What carries the argument

The load-bearing object is the infinite Jacobi operator $J=V^{1/2}CV^{1/2}$ obtained as the $N\to\infty$ limit of the symmetrised generalised capacitance matrix of the resonator chain, with off-diagonal bands $s(i)=v_{i-1}v_i s_{i-1}^{-1}(\ell_{i-1}\ell_i)^{-1/2}$ and diagonal entries $q(i)=v_i^2\ell_i^{-1}(s_{i-1}^{-1}+s_i^{-1})$. Because i.i.d. block sampling makes the resonator sequence a bi-infinite Markov chain, the shift group acts metrically transitively on $J$, yielding ergodicity of the spectrum and convergence of finite-size integrated densities of states. The second mechanism is the propagation-matrix formalism: each block has a $2\times2$ transfer matrix, and frequencies with $|\operatorname{tr}P_{B_d}(\lambda)|>2$ lie in a bandgap for that block, so in a hybridisation region at least one block type is gapped and the corresponding eigenmodes decay exponentially. That decay justifies replacing the array by a catalogue of finite meta-atoms—local sequences beginning and ending with the active block—whose defect eigenfrequencies, precomputed from small capacitance matrices, reproduce the hybridisation-region density of states.

What would settle it

Solve the full Helmholtz resonance problem (2.2) at a fixed small contrast $\delta>0$ for random block chains of increasing length $M$, and compare the empirical integrated density of states with the capacitance-matrix limit $N(J,\lambda)$; if the Wasserstein distance does not tend to zero as $M\to\infty$ and $\delta\to0$, the deterministic density is an artefact of the discrete approximation. A sharper test: for $\delta$ small enough that the $O(\delta)$ eigenvalue error is below the peak spacing, the exact subwavelength resonances should reproduce the same meta-atom peak positions in the hybridisation region.

Watch

Extended reading notes

Core claim

On its own terms, the central discovery is that the integrated density of states of an i.i.d. block-disordered one-dimensional resonator system converges almost surely as the number of blocks $M\to\infty$ to a non-random measure $N(J,\mathrm{d}\lambda)$, and the distribution function $\lambda\mapsto N(J,\lambda)$ is continuous: Theorem 3.9, imported from metric-transitivity theory. The spectrum is partitioned by the pass bands of the constituent blocks: frequencies gapped for every block carry no states; frequencies passed by every block form a smooth band; and frequencies passed by some but not all blocks form hybridisation regions with non-zero density concentrated on self-similar peaks that are only weakly smoothed as the system grows. The paper further claims that the peaks are the defect modes of finite local block arrangements, called meta-atoms, so the density of states can be predicted by enumerating meta-atoms, computing their defect eigenvalues once, and scanning the block sequence in linear time. The same meta-atom procedure is demonstrated for bound-length, hyperuniform chunk and softmax, and Fibonacci quasiperiodic sampling, where it typically performs as well as or better than for i.i.d. sampling.

Load-bearing premise

All theorems are proved for the discrete capacitance matrix, whose eigenvalues match the true subwavelength resonances only up to an $O(\delta)$ error, and the paper does not show this error is uniform as the number of blocks tends to infinity.

Editorial extensions

If this is right

  • For any single realization of a large random block chain, the empirical cumulative eigenvalue count approaches one fixed continuous non-random curve, so one large finite sample is a statistically representative proxy for the infinite system.
  • The density of states is zero on the intersection of all constituent blocks' bandgaps and positive in both shared pass bands and hybridisation regions, which classifies the observable spectrum using only the block propagation matrices.
  • The apparent fractal jumps in the empirical cumulative density are finite-size signatures: the infinite-limit integrated density is continuous, so the jagged structure is progressively smoothed as $M$ grows.
  • The meta-atom algorithm reconstructs the hybridisation-region cumulative density in $O(M)$ time (or $O(ML)$ when the catalogue is scaled with length), making the deterministic limit computable for very long chains.
  • The same linear-time meta-atom estimation works, and often converges faster, for bound-length, hyperuniform chunk and softmax, and Fibonacci quasiperiodic sampling, provided the meta-atom catalogue is tailored to the sampling rule.

Reading between the lines

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

  • A direct consequence of Theorem 2.1 is an open uniformity question the paper leaves implicit: the capacitance eigenvalues match the physical resonances only up to $O(\delta)$, and if that error is not uniform in $M$, the proven deterministic density of states describes the discrete tight-binding model rather than the Helmholtz resonators.
  • The decay mechanism is generic: in any one-dimensional aperiodic wave system where one constituent is opaque in a frequency band, the global density of states should be assemblable from a weighted catalogue of finite defect patterns, so the meta-atom idea could be tried on layered dielectrics, phononic chains, or non-Hermitian arrays.
  • The Section 4.2 caveat that the meta-atom estimator ignores edge effects means its advertised linear-time accuracy holds strictly away from boundaries; the authors argue this is acceptable because small arrays can be diagonalised directly, but it sets a precise domain of validity for the estimator.
  • A testable refinement suggested by the sampling comparison is to adapt the meta-atom set to the sampling rule, discarding impossible patterns and weighting by occurrence probability; this should remove the accuracy reversal observed for Fibonacci sequences at large meta-atom length.
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

4 major / 7 minor

Summary. The paper studies the density of states (DoS) of one-dimensional aperiodic block subwavelength resonator systems. It introduces a discrete capacitance model, forms the infinite-volume Jacobi operator, and uses Pastur's theory of metrically transitive operators to prove that, for i.i.d. random block sequences, the integrated density of states converges to a non-random, continuous function as the number of blocks goes to infinity. The authors also propose a tripartite spectral decomposition into bandgaps, shared pass bands, and hybridisation regions, and they develop a fast 'meta-atom' algorithm to estimate the DoS in the hybridisation region. Numerical experiments are presented for i.i.d., bound-length, hyperuniform, and quasiperiodic samplings.

Significance. If the claims hold, the paper gives a rigorous ergodic-theoretic foundation for the DoS of disordered subwavelength resonator arrays and offers a linear-time numerical method. The metric transitivity proof (Proposition 3.6) is clean, the application of Pastur's theorem is transparent, and the open code and detailed numerical comparisons are strengths. However, the physical relevance of the main theorem is limited by an unproven identification between continuous and discrete resonances, and the tripartite decomposition and the meta-atom algorithm are largely heuristic. The paper is therefore best seen as a rigorous study of the discrete capacitance model with suggestive numerical evidence for the physical system.

major comments (4)
  1. [Section 2.1, Theorem 2.1] The approximation ω_i(δ) = √δ λ_i + O(δ) is used to identify the physical resonant frequencies with the discrete eigenvalues, and the paper states 'we will often use λ_i and ω_i interchangeably'. However, all subsequent theorems and figures concern the eigenvalues λ_i of the generalised capacitance matrix (or of the Jacobi operator J). The O(δ) error is not shown to be uniform in the number of resonators N, so the empirical measure of ω_i/√δ could differ from that of λ_i in the thermodynamic limit M→∞ at fixed δ, and then δ→0. This gap directly affects the title's claim of a density of states for subwavelength resonator systems. Either provide a uniform (in N and i) error estimate or clearly restrict the claims to the discrete model.
  2. [Section 5, §§5.1–5.3] The paper extends the convergence and determinism results of Section 3 to dependent samplings. For bound-length sampling (§5.1) it asserts 'all the convergence results from Section 3 continue to hold also under this sampling' without proof; for softmax sampling (§5.2, Eq. (5.2)) the process is defined by a feedback rule and metric transitivity is not verified; for the Fibonacci tiling (§5.3) no probability space is even specified. Since Theorem 3.9 is quoted from Pastur for metrically transitive operators, each of these samplings requires explicit verification of the hypotheses. Without this, the universal determinism of the DoS for these cases is an assumption rather than a theorem.
  3. [Section 2.2, Figure 2; Section 4] The tripartite decomposition of the spectrum into shared pass band, bandgap, and hybridisation region is presented as a 'complete description'. The bandgap part is rigorously supported by Theorem 2.6, but the smoothness of the DoS in shared pass bands and the fractal-like behaviour in hybridisation regions are only demonstrated numerically. The analytical argument in §4.1 (Proposition 4.1) is a perturbation bound for a single eigenpair and does not imply the universal structure of the density of states. Please either provide formal statements (e.g., asymptotics or scaling laws for the IDS in the hybridisation region) or explicitly label this trichotomy as a numerical observation and adjust the abstract.
  4. [Section 4.2, Algorithm 1] The meta-atom algorithm is a central contribution, but it is presented without any theoretical error estimate. The Wasserstein-distance convergence shown in Figure 5 is empirical and depends on the chosen blocks and sampling parameters; no proved bound relates the output of Algorithm 1 to the infinite-volume IDS of Theorem 3.9. If the algorithm is intended as a heuristic numerical tool, the paper should state this; if it is intended as a 'universal estimate', a convergence theorem or error bound is needed.
minor comments (7)
  1. [Definition 2.5] The phrase 'from the left edge of the resonator x_i^L to the left edge of the following resonator x_i^L' is internally inconsistent; the second symbol should be x_{i+1}^L.
  2. [Figures 2–7] Labels like '10□18' appear to be rendering artifacts of negative exponents; please regenerate the figures so that all annotations are legible.
  3. [Remark 3.4] The truncation size is denoted by N (with (2N+1) resonators), while N was earlier used for the total number of resonators in a finite system. This overloading is confusing and should be disambiguated.
  4. [References [4], [6]] The core results rely heavily on the companion preprint [4] and the forthcoming book [6]. Please state the dependence explicitly and, where possible, include the relevant statements in the text or make the references available.
  5. [Appendix B, Eq. (B.1)] The notation 'C_α := V C_α' reuses the same symbol for the matrix and its generalised version; a distinct notation would improve clarity.
  6. [Figure 3 caption] The caption says 'M = 105 realisations', which presumably means 10^5 realisations; please correct.
  7. [Abstract and §4] The term 'fractal-like' is used without a precise definition; since the paper itself notes that hybridisation smooths the density, please clarify the intended meaning (e.g., self-similarity of peaks at a fixed resolution).

Circularity Check

0 steps flagged · score 2.0 of 10

No direct circularity: the DoS convergence rests on Pastur's external ergodic theorem, and the meta-atom estimates are non-fitted approximations; self-citations are present but not load-bearing.

full rationale

The paper's central convergence claim, Theorem 3.9, is quoted from Pastur's external monograph [27]; it is not derived from the authors' prior work. The only inputs are the i.i.d./Markov block-sampling construction and the verification that the induced resonator sequences give a shift-invariant Jacobi operator. Proposition 3.6 checks the metric-transitivity hypothesis directly, so the application of [27] is a genuine import of an external theorem rather than a restatement of the desired conclusion. The meta-atom algorithm likewise does not fit parameters to the target density of states: it enumerates local block arrangements, computes their spectra from the same generalized capacitance model, and compares the resulting CDF to the full empirical CDF via the Wasserstein distance, with Proposition 4.1 providing an a priori perturbation estimate. The tripartite spectral classification uses Theorem 2.6 quoted from [4], a self-citation, but the DoS convergence does not depend on that theorem, and the theorem is a checkable mathematical statement rather than a fitted output. The only substantive caveat is the identification of lambda_i with omega_i via Theorem 2.1's O(delta) error, whose uniformity in N is not established; this is an approximation gap and a physical-modeling risk, not a circular definition. No equation in the paper reduces to its own output by construction, so there is no significant circularity.

Assumptions & free parameters 4 free parameters · 6 assumptions · 1 invented entities

Free parameters are the hyperparameters of the meta-atom algorithm (R, L, P) and the unreported softmax temperature beta; none of these are fitted to the target density of states, so they do not constitute circular fitting, but they are user-chosen values. The axioms show the paper's dependence on three external inputs: the capacitance matrix approximation, Pastur's ergodic theorems (including the unverified continuity hypothesis), and the heuristic spectral trichotomy. The meta-atom is the only invented construct and it is a computational ansatz with internal numerical support only.

free parameters (4)
  • R: number of attached single-resonator buffers in meta-atom evaluation = 4
    Chosen by hand in Section 4.2; justified as sufficient because eigenmodes decay exponentially in the single-resonator bandgap, but no convergence study in R is shown.
  • L: maximum meta-atom length = varied up to 14
    User-chosen truncation of the meta-atom set M^P_L; the paper shows accuracy improves with L, but the choice is arbitrary and convergence is only empirically demonstrated in Figures 5 and 8.
  • P: maximum number of single resonators in a meta-atom = scaled as L/2
    User-chosen truncation parameter in M^P_L, set together with L; no principled rule is given for the scaling.
  • beta: softmax sampling temperature = not stated
    Softmax sampling in equation (5.2) depends on beta, which controls the strength of occurrence-count regularization; the beta values used in Figures 7(c), 8, and 9 are not reported, making the dependent-sampling results hard to reproduce.
assumptions (6)
  • domain assumption The generalized capacitance matrix C=VC accurately describes the N subwavelength resonant frequencies in leading order as delta tends to 0 (Theorem 2.1 of [21]), and this approximation is stable as N tends to infinity.
    Invoked in Section 2.1 and used throughout; all spectral analysis is on the discrete matrix rather than the original Helmholtz problem, and no uniform-in-N error estimate is provided.
  • standard math Pastur's ergodic theorems for metrically transitive operators apply to the Jacobi operator J(mu) built from the resonator sequence mu (Theorem 3.9).
    The paper quotes [27, Theorems 3.2 and 4.5] as Theorems 3.7 and 3.9; the proof reduces to verifying metric transitivity, which the paper does in Proposition 3.6.
  • ad hoc to paper The integrated density of states lambda maps to N(J, lambda) is continuous for the block-disordered Jacobi operators considered (Theorem 3.9(iii)).
    Continuity is quoted from Pastur, but the specific hypothesis guaranteeing absence of atoms in the IDS is not verified for this operator class; if that condition failed, the abstract's 'continuous' claim would not be established.
  • domain assumption With i.i.d. sampling, any frequency in the pass band of at least one constituent block is approached by eigenvalues of the full random system, so the spectral trichotomy (shared pass band, bandgap, hybridisation region) is exhaustive.
    The 'only if' part of the bandgap characterization in Section 2 is justified by a probabilistic argument about arbitrarily long same-block runs, but no rigorous spectral proof is given; the trichotomy underlies the definition of hybridisation regions used throughout Section 4.
  • standard math Eigenmodes at frequencies in the single-resonator bandgap decay exponentially through single-resonator blocks, so local dimer arrangements are only weakly coupled.
    This follows from the transfer and propagation matrix formalism in Section 2.4 and is the basis for the meta-atom approximation in Section 4.1.
  • domain assumption The finite capacitance matrix eigenvalue count converges to the IDS of the infinite Jacobi operator as N tends to infinity, with edge effects negligible.
    Stated in Section 3.2 and Remark 3.4; the identification between the eCDF of C and the K-truncation IDS N_K is asserted but not proven beyond noting that edge effects vanish.
invented entities (1)
  • Meta-atoms: finite block sequences constrained by M^P_L used to compute defect modes
    purpose: Enable fast prediction of the density of states in hybridisation regions by enumerating local arrangements and weighing them by occurrence probability.
    The meta-atom is a computational construct, not a physical entity. Its predictions are checked numerically against full diagonalization of the same discrete model within the paper, but no experimental or otherwise external validation is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Universal estimates for the density of states for aperiodic block subwavelength resonator systems." pith.science (2026). https://pith.science/paper/5ANOISFP

@misc{pith2026250516677,
  author       = {Pith},
  title        = {Pith review of: Universal estimates for the density of states for aperiodic block subwavelength resonator systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5ANOISFP}},
  note         = {Machine review of arXiv:2505.16677}
}
abstract

We consider the spectral properties of aperiodic block subwavelength resonator systems in one dimension, with a primary focus on the density of states. We prove that for random block configurations, as the number of blocks $M\to \infty$, the integrated density of states converges to a non-random, continuous function. We show both analytically and numerically that the density of states exhibits a tripartite decomposition: it vanishes identically within bandgaps; it forms smooth, band-like distributions in shared pass bands (a consequence of constructive eigenmode interactions); and, most notably, it exhibits a distinct fractal-like character in hybridisation regions. We demonstrate that this fractal-like behaviour stems from the limited interaction between eigenmodes within these hybridisation regions. Capitalising on this insight, we introduce an efficient meta-atom approach that enables rapid and accurate prediction of the density of states in these hybridisation regions. This approach is shown to extend to systems with quasiperiodic and hyperuniform arrangements of blocks.

Figures

Figures reproduced from arXiv: 2505.16677 by the authors.

Figure 1
Figure 1. A block disordered system consisting of two single resonator blocks B1 and a dimer block B2 arranged in a chain given by the sequence χ = (1, 2, 1). It thus consists of M = 3 blocks and N = 4 resonators D1, . . . , D4 in total. that denote the wave speed, length, and spacing of each constituent resonator. Here, len(Bj ) denotes the total number of resonators contained within the block Bj . We will often abuse notati… view at source ↗
Figure 2
Figure 2. Left: Comparison of maximal block propagation matrix eigen￾value |ξ2(λ)| for both blocks with Thouless ratios g(λi) of the entire random block disordered system consisting of these block (each colored vertical line corresponds to the Thouless ratio g(λi) of the eigenvalue λi). Where there are no such lines, the density of eigenvalues is zero. Right: Cumulative density func￾tion (CDF) of the total system, laid atop t… view at source ↗
Figure 3
Figure 3. Convergence of empirical cumulative density functions under increasing system size. We consider random block disordered systems consisting of single resonator and dimer blocks as in Ex￾ample 2.2 sampled with either equal density (psingle = pdimer = 1/2) or low dimer density (pdimer = 1/10). We calculate the eCDF for large arrays (M = 213) and compare this to the eCDF of smaller resonator arrays (M = 2p , p = 2, . . … view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Densities of states for block disordered systems (M = 105 ) with varying dimer density, together with various dimer defect modes. We can see that in both cases (but especially in the low dimer density case) the peaks in the density of states closely correspond to the d…
Figure 5
Figure 5. Figure 5: Convergence of the meta-atom estimate using Algorithm 1 to the empirical CDF of a large system as the meta-atom length L and single resonator amount P is increased. Already for small numbers of L and P, the estimate agrees extremely well with the empirical CDF. The gre…
Figure 6
Figure 6. Figure 6: As we decrease the first spacing s1 of the dimer blocks, the corresponding upper band at (2/s1, 2/s1+1) gets pushed further into the bandgap (1, ∞) of the single resonator block. This causes weaker hybridisation which in turn increases the accuracy of the meta-atom est…
Figure 7
Figure 7. Figure 7: Densities of states for a variety of block disordered systems with dependent sampling, together with various dimer defect modes. All of them consist of single resonator and dimer blocks as described in Example 2.2. We can see that the densities are dominated by just a …
Figure 8
Figure 8. Figure 8: Wasserstein distance between the empirical cumulative density function and the meta-atom estimate for increasing meta-atom length L (the amount of permissible single resonators is scaled as L/2) for block disordered systems constructed according variety of block sampli…
Figure 9
Figure 9. Figure 9: Fourier transform Kb(k) of the autocovariance for a variety of sampling methods. We can see that for both the hyperuniform chunk sampling and the softmax sampling, the Fourier trans￾form Kb(k) → 0 as |k| → 0. The autocovariance is calculated empirically over sequences …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

10 extracted references · 9 canonical work pages

  1. [28]

    Computation and visualization of photonic quasicrystal spectra via Bloch’s theorem

    Rodriguez Alejandro W, McCauley Alexander P, A vnielYehuda, andJohnson Steven G, “Computation and visualization of photonic quasicrystal spectra via Bloch’s theorem”, in:Physical Review B77.10 (2008), p. 104201

    work page 2008

  2. [29]

    Some electronic properties of a one-dimensional crystal model

    Saxon D. S. andHutner R. A., “Some electronic properties of a one-dimensional crystal model”, in:Philips J. Res.4 (1949), pp. 81–122

    work page 1949

  3. [30]

    Almost periodic functions and partial differential operators

    Shubin Mikhail Aleksandrovich, “Almost periodic functions and partial differential operators”, in:Russian Mathematical Surveys33.2 (1978), p. 1

    work page 1978

  4. [31]

    Almost periodic Schrödinger operators: a review

    Simon Barry, “Almost periodic Schrödinger operators: a review”, in:Advances in Applied Mathematics3.4 (1982), pp. 463–490

    work page 1982

  5. [32]

    Electrons in Disordered Systems and the Theory of Localization

    Thouless D.J., “Electrons in Disordered Systems and the Theory of Localization”, in: Phys. Rep.13.3 (1974), pp. 93–142

    work page 1974

  6. [33]

    Hyperuniform states of matter

    Torquato Salvatore, “Hyperuniform states of matter”, in:Phys. Rep.745 (2018), pp. 1–95

    work page 2018

  7. [34]

    Light in correlated disordered media

    Vynck Kevin, Pierrat Romain, Carminati Rémi, Froufe-Pérez Luis S., Schef- fold Frank, Sapienza Riccardo, Vignolini Silvia, and Sáenz Juan José, “Light in correlated disordered media”, in:Rev. Mod. Phys.95.4 (2023), p. 045003

    work page 2023

  8. [35]

    Engineered disorder in photonics

    Yu Sunkyu, Qiu Cheng-Wei, Chong Yidong, Torquato Salvatore, and Park Namkyoo, “Engineered disorder in photonics”, in:Nature Reviews Materials6.3 (2021), pp. 226–243

    work page 2021

Show all 10 references
  1. [36]

    Crystalline Metamaterials for Topological Properties at Subwavelength Scales

    Yves Simon, FleuryRomain, BerthelotThomas, Fink Mathias, LemoultFabrice, and Lerosey Geoffroy, “Crystalline Metamaterials for Topological Properties at Subwavelength Scales”, in:Nat. Commun.8.1 (2017), p. 16023

    2017

  2. [37]

    A Review on Non-Hermitian Skin Effect

    Zhang Xiujuan, Zhang Tian, Lu Ming-Hui, and Chen Yan-Feng, “A Review on Non-Hermitian Skin Effect”, in:Advances in Physics: X7.1 (2022), p. 2109431,doi: 10.1080/23746149.2022.2109431. Habib Ammari ETH Zürich, Department of Mathematics, Rämistrasse 101, 8092 Zürich, Switzerland...

    2022

Pith tools

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