Pith. sign in

REVIEW 2 major objections 4 minor 96 references

Universal Methods for Nonlinear Spectral Problems

T0 review · 2 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

Pith's one-line read This paper claims the first general, convergent computational method for the spectra and pseudospectra of nonlinear spectral problems, applicable to any gap-continuous pencil.

desk verdict The SCI classification and convergent algorithms for gap-continuous nonlinear pencils are real and mostly correct, but the 'universal' scope is narrower than claimed: the core-basis condition is an extra hypothesis, and the differential examples rely on an unproved relaxation to z-dependent bases. read the letter →

arxiv 2504.17012 v1 pith:X252PWKZ submitted 2025-04-23 math.NA cs.NAmath.SP

classification math.NAcs.NAmath.SP MSC 35P3046N4047A1047J1065Hxx65J1065N30
keywords nonlinearspectralproblemspollutioninvisibilityinjectionmoduluspseudospectrasolvabilitycomplexityindexgaptopologyoperatorpencils
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 aims to settle a basic question: when the spectral parameter enters an operator pencil nonlinearly, can the spectrum and pseudospectra (regions where the inverse of the pencil is nearly singular) be computed without spectral pollution from spurious points or spectral invisibility from missed true points? The authors propose two universal algorithms and prove that they converge for every nonlinear pencil that is continuous in the gap topology on operator graphs, a very weak continuity assumption. Under the Solvability Complexity Index hierarchy, spectra require three successive limits when only matrix entries are known and two when quadratic data about $T(z)^*T(z)$ and $T(z)T(z)^*$ are available; pseudospectra require two and one, respectively, and the paper proves that no method can use fewer limits. A separate theorem shows that Hermiticity does not lower these counts, in contrast to the linear case. Worked examples include nonlinear shifts, Klein–Gordon pencils, acoustic wave equations, fractional viscoelastic beams, and a delayed predator–prey model.

What carries the argument

The carrying object is the nonlinear injection modulus and its monotone finite truncations: $\gamma(z,T)=\min\{\sigma_{\inf}(T(z)),\sigma_{\inf}(T(z)^*)\}$, where $\sigma_{\inf}(A)=\inf_{\|x\|=1}\|Ax\|$ measures how far an operator is from failing to be injectively bounded below. $\gamma$ is continuous with respect to the gap metric on operator graphs, and the finite rectangular blocks $Q_{n_1}T(z)P_{n_2}^*$ give computable lower bounds whose limit is the modulus itself, while the unbounded truncations $T(z)P_{n_2}^*$ give upper bounds. The convergence of these approximations is monotone, and the Attouch–Wets metric on relatively closed subsets of $U$, which weighs accuracy by distance to the boundary of $U$, turns the informal goals of no pollution and no invisibility into a precise convergence theorem.

What would settle it

Run Algorithm 1 on any pencil in the stated class and compare its output to the true $Sp_\epsilon(T)$ under the Attouch–Wets metric; any pencil where the limit differs would refute the convergence claim. A sharper falsifier targets the relaxed-basis assertion: for the acoustic wave pencil on the half-line with spectrum $\{z:\mathrm{Im}(z)\ge 0\}$, compute $\gamma_{n_2}(z,T)$ using the paper's $z$-dependent orthonormal bases at a fixed $z$ with $\mathrm{Im}(z)>0$; if a single such value falls strictly below $\gamma(z,T)=\|T(z)^{-1}\|^{-1}$, the convergence-from-above step fails and those examples are not covered by the theorems.

Watch

Extended reading notes

Core claim

The central claim is that the nonlinear spectral problem is solvable in full generality by one mechanism: the nonlinear injection modulus $\gamma(z,T)=\min\{\sigma_{\inf}(T(z)),\sigma_{\inf}(T(z)^*)\}$, which equals $\|T(z)^{-1}\|^{-1}$ when $T(z)$ is invertible and vanishes exactly on the spectrum. Truncated versions of this modulus, computed from finite rectangular matrix blocks, converge monotonically as the truncation size grows, from below for the finite-block approximations and from above for the unbounded truncations. Assembling these truncated moduli over dense grids yields Algorithms 1 and 2, which the paper proves converge to the pseudospectra and the spectrum in the Attouch–Wets metric for every pencil in the class $\Omega^U_{\mathrm{NL}}$, thereby excluding both pollution and invisibility. The same theorems give matching lower bounds in the Solvability Complexity Index hierarchy, showing optimality in any model of computation, and extend the lower bounds to Hermitian pencils.

Load-bearing premise

The load-bearing premise is that the fixed basis vectors form a core of $T(z)$ and of $T(z)^*$ at every point $z$; if that core property fails, the monotone convergence of the truncated injection moduli from above, which is what excludes invisibility, can break down—and for several differential-operator examples the authors relax this to $z$-dependent bases without giving a proof that the convergence theorems still hold.

Editorial extensions

If this is right

  • For any gap-continuous nonlinear pencil, pseudospectra can be approximated with two limits under $\Lambda_1$ and one under $\Lambda_2$, while spectra need three and two; these counts are optimal.
  • Because the $\Lambda_2$ approximations converge monotonically from above, their output sets stay inside the true pseudospectra, so interval-arithmetic variants produce rigorous enclosures suitable for computer-assisted proofs.
  • Spectral pollution and invisibility are excluded by construction in the limit, not by problem-specific tuning, so the method applies equally to discrete, differential, and lattice-type pencils.
  • Hermiticity does not make nonlinear spectral problems easier: the required number of limits is unchanged, so the extra cost over the linear case is intrinsic to the nonlinearity.
  • The singular vectors used to compute the moduli also give approximations of pseudoeigenfunctions, and the resulting pseudospectral bounds are sharper than numerical-range-type estimates used in contour-based evolution solvers.

Reading between the lines

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

  • I infer that the same monotone-truncation construction should carry over to pencils parametrized over manifolds or to operator-valued functions with an additional parameter, since the planar geometry enters only through the Attouch–Wets metric on subsets of $U$.
  • The unproved $z$-dependent-basis relaxation is the main internal risk: the three differential-operator demonstrations are not fully covered by Theorems 3.1 and 3.2 unless that relaxed assumption is proved, and a focused proof or counterexample would settle whether those examples are theorems or merely numerical evidence.
  • I expect that feeding the sharper computed pseudospectra into contour-integration solvers for time-fractional or wave-type equations would improve convergence estimates compared with numerical-range bounds, but that effect is not quantified in this paper.
  • A practical extension would be to compute only selected parts of the spectrum, such as eigenvalues inside a contour, by restricting the grids and localizing the injection-modulus search, since the current algorithms compute global pseudospectra first.
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

2 major / 4 minor

Summary. The paper develops a general computational framework for spectra and pseudospectra of nonlinear operator pencils T: U -> C(H1,H2) that are continuous in the gap topology. It introduces nonlinear injection moduli, proves their continuity, and uses monotone finite-dimensional truncations to construct two algorithms: Algorithm 1 computes pseudospectra with two (or one) limits, and Algorithm 2 computes spectra with three (or two) limits, depending on whether the evaluation set is Lambda_1 or Lambda_2. The main theoretical results are the SCI classifications in Theorems 3.1 and 3.2, including lower bounds showing optimality and the claim that Hermiticity does not reduce the number of limits for nonlinear problems. The paper also presents numerical examples for nonlinear shifts, Klein-Gordon equations, wave equations with acoustic boundary conditions, time-fractional beam equations, and a delayed reaction-diffusion system.

Significance. If the results stand, this is a substantial contribution to computational spectral theory. The central chain is sound: Lemma 3.1 characterizes the spectrum as the zero set of the injection modulus, Lemma 3.2 proves monotone convergence of the truncated moduli, Propositions 3.3 and 3.4 establish Attouch-Wets convergence of the algorithms, and Theorem 3.1 gives matching SCI upper and lower bounds. The constructions are parameter-free, do not use fitted constants, and the lower-bound reductions are explicit and machine-checkable in principle. The numerical examples are extensive and illustrate both the strengths of the method and the failure of naive truncation. However, the actual scope of the theorems is narrower than the abstract claims: the core-basis conditions in Definition 2.4 are extra hypotheses that are not implied by gap continuity, and the relaxed z-dependent basis used in Section 4.3 is asserted without proof. These issues are load-bearing because they affect the no-invisibility guarantee.

major comments (2)
  1. [Abstract and §1; Definition 2.4 and Lemma 3.2(ii)] The abstract and the introduction state that the method requires only continuity with respect to the gap metric, but Definition 2.4 adds two further conditions: for every z in U, span{e_n} must be a core of T(z) and span{e_hat_n} must be a core of T(z)*. These conditions are not consequences of gap continuity. For example, T(z)=M_{1/(x-z)} on L^2(0,1) with U=(0,1) is gap-continuous, yet the intersection of the domains over all z is {0}, so no fixed orthonormal basis can have finite spans forming cores for every z. Lemma 3.2(ii) relies on both core conditions; if they fail, gamma_{n2}(z,T) can converge from above to a value strictly larger than gamma(z,T), and Algorithm 2 can miss true spectral points. Thus Theorem 3.1 classifies only the narrower class Omega_U_NL, and the claim of applicability to all gap-continuous pencils should be corrected, or the theorem extended to a properly stated weaker hypothesis.
  2. [§4.3–4.5] Section 4.3 introduces a z-dependent orthonormal basis obtained by Gram-Schmidt and states without proof that 'the convergence results in Section 3.3 continue to hold under this relaxed assumption.' The theorems require a fixed basis whose spans are cores of both T(z) and T(z)*. In the wave-equation example the basis functions do not lie in the domain of T(z), the modified basis depends on z, and no core condition is verified for T(z)*. If the second term of gamma_{n2}(z,T) fails to converge down to gamma(z,T), spectral invisibility can occur. Sections 4.4 and 4.5 also do not verify the core condition for the bases used. Consequently, the differential-operator demonstrations are not consequences of Theorems 3.1 and 3.2 as written. The authors should either provide a proof for a relaxed-core theorem covering their z-dependent and Gram-Schmidt constructions, or explicitly present these examples as numerical demonstrations under an unverified heuristic assumption.
minor comments (4)
  1. [Throughout] There are several typos, including 'Pseudosepctra' in the caption of Figure 1 and 'the prof if' in the proof of Proposition 3.4.
  2. [§2.3] The extension of the Attouch-Wets metric to the empty set is described by a convergence convention rather than by an explicit metric. For a fully rigorous SCI formulation, the empty set should be incorporated into the metric space with a defined distance, or the convention should be justified as a limiting metric.
  3. [References] Several references are incomplete or lack publication details, including [9], [10], [11], [13], [24], [39], and [47]; for example, [9] lists only volume and pages without a journal or article title.
  4. [§4.1–4.2] The notation for the truncated injection modulus alternates between gamma_{n2}(z,T) and gamma_n(z,T), and in Figure 4 the legend uses both; standardizing this notation would improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the spectral algorithms are derived from an explicit injection-modulus identity with proven convergence, and the cited prior hardness results are independent support rather than an input-output loop.

full rationale

The derivation is self-contained and non-circular. The central reduction is Eq. (3.1), Sp(T)={z in U: gamma(z,T)=0} and Sp_epsilon(T)=Cl_U({z: gamma(z,T)<epsilon}), which follows from Lemma 3.1 proved in the paper (injection modulus equals the inverse norm when boundedly invertible, and min{sigma_inf,sigma_inf*}=0 otherwise). Algorithms 1 and 2 then operate on finite-dimensional singular values gamma_{n2,n1}(z,T) computed from Lambda_1/Lambda_2 matrix elements; the convergence statements in Propositions 3.3 and 3.4 are proved by monotone approximation and continuity of gamma, not by assuming the spectrum. No fitted parameter, calibration set, or predicted quantity is defined in terms of the output, and the output sets are not used as inputs to the algorithm. The lower bounds in Theorem 3.1 and Proposition 3.5 do cite the authors' earlier papers [6], [26], and [28], but those are published, parameter-free hardness theorems for linear and decision problems whose assumptions do not include the nonlinear spectral problem, so they are external support rather than a self-citation loop. The one notable gap is in Section 4.3, where the authors state 'The convergence results in Section 3.3 continue to hold under this relaxed assumption' after introducing z-dependent Gram-Schmidt bases; this is an unproved extension, and Definition 2.4's core condition is an extra hypothesis, but this is a correctness or assumption gap in the examples, not an equivalence between output and input, and it does not make the core derivation circular. Hence no circularity; score 0.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The central derivation is parameter-free: Equations (3.1) to (3.5) define spectra and pseudospectra in terms of the injection modulus and prove monotone convergence of its finite truncations, with no fitted constants; the 1/n2 factors are convergence slacks, and example parameters (d1, d2, tau, r, r1, r2) are inputs taken from the cited biology and model literature. The load-bearing premises are the gap-continuity and core class definition, the grid and evaluation-set model, and the imported hardness theorems (mostly self-cited but published and parameter-free). No new entities are postulated.

assumptions (6)
  • standard math Kato's gap-metric theory of closed operators, including generalized convergence and Theorem 2.1 properties.
    Invoked in Definition 2.4 and Proposition 3.1 to connect gap convergence with resolvent and injection-modulus behaviour; standard operator theory, not re-proven.
  • domain assumption For each z in U, span{e_n} is a core of T(z) and span{e_hat_n} is a core of T(z)* (Definition 2.4).
    Necessary for Lemma 3.2(ii) monotone convergence from above of gamma_{n2} to gamma; if it fails, the no-invisibility guarantee does not follow. Also the practical bottleneck in the examples.
  • standard math Prior hardness theorems: linear spectrum and pseudospectrum lower bounds of [6, 46] and the Delta^G_3 bound for Xi_{2,Q} of [28].
    Imported for the lower bounds in Theorem 3.1 and Proposition 3.5; published, parameter-free theorems, but authored by the present authors, so treated as independent support with a small self-citation burden.
  • domain assumption Dense rational grids G_n with the stated density property (Section 2.4) and exact evaluation of matrix elements at grid points.
    Convergence in Propositions 3.3 and 3.4 uses grid density; finite-precision effects are handled informally in Remark 9 and are not incorporated into the statements.
  • domain assumption Unproved example facts: Sp(T) = {z : Im(z) >= 0} for the acoustic problem (4.3); essential spectrum of H0 is [0,1] union [3,4] for the Klein-Gordon example (4.2).
    Stated without derivation or citation; these support the illustrative correctness of the examples, not the central theorems.
  • ad hoc to paper Convergence of the algorithms extends to z-dependent orthonormal bases built by Gram-Schmidt (Section 4.3).
    Asserted as 'the convergence results in Section 3.3 continue to hold under this relaxed assumption', with no proof given; three example families depend on it.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Universal Methods for Nonlinear Spectral Problems." pith.science (2026). https://pith.science/paper/X252PWKZ

@misc{pith2026250417012,
  author       = {Pith},
  title        = {Pith review of: Universal Methods for Nonlinear Spectral Problems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/X252PWKZ}},
  note         = {Machine review of arXiv:2504.17012}
}
read the original abstract

Nonlinear spectral problems arise across a range of fields, including mechanical vibrations, fluid-solid interactions, and photonic crystals. Discretizing infinite-dimensional nonlinear spectral problems often introduces significant computational challenges, particularly spectral pollution and invisibility, which can distort or obscure the true underlying spectrum. We present the first general, convergent computational method for computing the spectra and pseudospectra of nonlinear spectral problems. Our approach uses new results on nonlinear injection moduli and requires only minimal continuity assumptions: specifically, continuity with respect to the gap metric on operator graphs, making it applicable to a broad class of problems. We use the Solvability Complexity Index (SCI) hierarchy, which has recently been used to resolve the classical linear problem, to systematically classify the computational complexity of nonlinear spectral problems. Our results establish the optimality of the method and reveal that Hermiticity does not necessarily simplify the computational complexity of these nonlinear problems. Comprehensive examples -- including nonlinear shifts, Klein--Gordon equations, wave equations with acoustic boundary conditions, time-fractional beam equations, and biologically inspired delay differential equations -- demonstrate the robustness, accuracy, and broad applicability of our methodology.

Figures

Figures reproduced from arXiv: 2504.17012 by the authors.

Figure 1
Figure 1. Pseudosepctra of 40 × 40 truncation of the nonlinear shift S − f(z)S ∗ for f(z) = sin(4z)(|z| 2 + 1). The color corresponds to a logarithmic grid of ϵ values. The spectrum of the truncated problem is {πm/4 : m ∈ Z}, which is completely different to the spectrum Sp(S − f(z)S ∗ ) = {z ∈ C : |f(z)| = 1} [PITH_FULL_IMAGE:figures/full_fig_p019_1.png] view at source ↗
Figure 2
Figure 2. Approximation of pseudospectra for the nonlinear shift example using Algorithm 1 [PITH_FULL_IMAGE:figures/full_fig_p019_2.png] view at source ↗
Figure 3
Figure 3. Output of Algorithm 1 for the Klein–Gordon spectral problem. The green and red dots [PITH_FULL_IMAGE:figures/full_fig_p021_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Convergence of local minima of γn(z, T) (essentially an adaptation of Algorithm 2) for three representative eigenvalues in the discrete spectrum. The legend shows the (rounded) approximate location of the eigenvalues. The plateauing of the curves around 10−15 is due to…
Figure 5
Figure 5. Figure 5: Left: Eigenvalues of the quadratic eigenvalue problem after domain truncation fol [PITH_FULL_IMAGE:figures/full_fig_p022_5.png]
Figure 6
Figure 6. Figure 6: Convergence of Algorithm 1 (as we move from left to right) for the wave equation [PITH_FULL_IMAGE:figures/full_fig_p023_6.png]
Figure 7
Figure 7. Figure 7: Convergence of pseudoeigenfunctions for z = 2π + i. orthonormal basis {fn}∞ n=1. This resulting orthonormal basis is dependent on z, illustrating a generalisation of our convergence results to cases where the bases used to represent the nonlinear pencil (in H1 and H2) …
Figure 8
Figure 8. Figure 8: Comparison of pseudospectra computed using Algorithm 1 (left) and the numerical [PITH_FULL_IMAGE:figures/full_fig_p025_8.png]
Figure 9
Figure 9. Figure 9: Similar to Fig. 8, but for [PITH_FULL_IMAGE:figures/full_fig_p025_9.png]
Figure 10
Figure 10. Figure 10: Similar to Fig. 8, but for [PITH_FULL_IMAGE:figures/full_fig_p026_10.png]
Figure 11
Figure 11. Figure 11: Steady-state solutions of the diffusive Lotka–Volterra predator-prey system with [PITH_FULL_IMAGE:figures/full_fig_p026_11.png]
Figure 12
Figure 12. Figure 12: Pseudospectra of the linearised predator-prey system, computed using Algorithm 1. [PITH_FULL_IMAGE:figures/full_fig_p027_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

96 extracted references · 77 canonical work pages

  1. [1]

    Adcock and A

    B. Adcock and A. C. Hansen. Compressive Imaging: Structure, Sampling, Learning . Cambridge University Press, Cambridge, 2021

  2. [2]

    Arendt, C

    W. Arendt, C. J. K. Batty, M. Hieber, and F. Neubrander. Cauchy Problems. In Vector-valued Laplace Transforms and Cauchy Problems , pages 109–240. Springer Basel, 2001

  3. [3]

    Baleanu, K

    D. Baleanu, K. Diethelm, E. Scalas, and J. J. Trujillo. Fractional calculus: models and numerical methods, volume 3. World Scientific, 2012

  4. [4]

    Bastounis, A

    A. Bastounis, A. C. Hansen, and V. Vlaˇ ci´ c. The extended Smale’s 9th problem – On computational barriers and paradoxes in estimation, regularisation, computer-assisted proofs and learning. arXiv preprint arXiv:2110.15734, 2021

  5. [5]

    G. Beer. Topologies on Closed and Closed Convex Sets , volume 268 of Mathematics and its Appli- cations. Springer Netherlands, 1993

  6. [6]

    Ben-Artzi, M

    J. Ben-Artzi, M. J. Colbrook, A. C. Hansen, O. Nevanlinna, and M. Seidel. Computing spec- tra - On the solvability complexity index hierarchy and towers of algorithms. arXiv preprint arXiv:1508.03280, 2020

  7. [7]

    Ben-Artzi, M

    J. Ben-Artzi, M. Marletta, and F. R¨ osler. Computing scattering resonances.Journal of the European Mathematical Society, 25(9):3633–3663, 2022

  8. [8]

    Ben-Artzi, M

    J. Ben-Artzi, M. Marletta, and F. R¨ osler. Computing the sound of the sea in a seashell.Foundations of Computational Mathematics , 22(3):697–731, 2022

Show all 96 references
  1. [9]

    Betcke, N

    T. Betcke, N. J. Higham, V. Mehrmann, C. Schr¨ oder, and F. Tisseur. NLEVP: A collection of nonlinear eigenvalue problems. 39(2):1–28, 2013

  2. [10]

    Betcke and L

    T. Betcke and L. N. Trefethen. Reviving the method of particular solutions. 47(3):469–491, 2005

  3. [11]

    Bindel and M

    D. Bindel and M. Zworski. Theory and computation of resonances in 1d scattering. 2006

  4. [12]

    L. Blum, F. Cucker, M. Shub, and S. Smale. Complexity and Real Computation . Springer New York, Secaucus, NJ, USA, 1998

  5. [13]

    Botchev, G

    M. Botchev, G. Sleijpen, and A. Sopaheluwakan. An SVD-approach to Jacobi–Davidson solution of nonlinear Helmholtz eigenvalue problems. 431(3-4):427–440, 2009

  6. [14]

    B¨ ottcher

    A. B¨ ottcher. Pseudospectra and singular values of large convolution operators. 6(3):267–301, Sept. 1994

  7. [15]

    B¨ ottcher

    A. B¨ ottcher. Infinite matrices and projection methods. In Lectures on operator theory and its applications (Waterloo, ON, 1994) , volume 3 of Fields Institute Monographs, pages 1–72. American Mathematical Society, 1996

  8. [16]

    B¨ ottcher and B

    A. B¨ ottcher and B. Silbermann. The finite section method for Toeplitz operators on the quarter- plane with piecewise continuous symbols. Math. Nachr., 110:279–291, 1983

  9. [17]

    B¨ ottcher and B

    A. B¨ ottcher and B. Silbermann. Introduction to Large Truncated Toeplitz Matrices . Universitext. Springer-Verlag, New York, 1999

  10. [18]

    B¨ ottcher and B

    A. B¨ ottcher and B. Silbermann. Analysis of Toeplitz Operators . Springer Monographs in Mathe- matics. Springer Berlin Heidelberg, second edition, 2006

  11. [19]

    B¨ ottcher and I

    A. B¨ ottcher and I. M. Spitkovsky. A gentle guide to the basics of two projections theory. Linear Algebra and its Applications , 432(6):1412–1459, 2010

  12. [20]

    N. P. Brown. AF embeddings and the numerical computation of spectra in irrational rotation algebras. Numerical Functional Analysis and Optimization , 27(5-6):517–528, 2006

  13. [21]

    N. P. Brown. Invariant means and finite representation theory of C*-algebras. Memoirs of the American Mathematical Society, 184(865), 2006

  14. [22]

    N. P. Brown. Quasi-diagonality and the finite section method. Mathematics of Computation , 76(257):339–361, 2007. 28

  15. [23]

    M. Caputo. Linear models of dissipation whose Q is almost frequency independent-II. Geophysical Journal International, 13(5):529–539, 1967

  16. [24]

    Chaitin-Chatelin and M

    F. Chaitin-Chatelin and M. B. van Gijzen. Analysis of parameterized quadratic eigenvalue problems in computational acoustics with homotopic deviation theory. 13(6):487–512, 2006

  17. [25]

    Chandler-Wilde, R

    S. Chandler-Wilde, R. Chonchaiya, and M. Lindner. On spectral inclusion sets and computing the spectra and pseudospectra of bounded linear operators. Journal of Spectral Theory, 14(2):719–804, 2024

  18. [26]

    M. J. Colbrook. The Foundations of Infinite-dimensional Spectral Computations . PhD thesis, University of Cambridge, 2020

  19. [27]

    M. J. Colbrook. Computing semigroups with error control. SIAM Journal on Numerical Analysis , 60(1):396–422, 2022

  20. [28]

    M. J. Colbrook. On the computation of geometric features of spectra of linear operators on Hilbert spaces. Foundations of Computational Mathematics , pages 1–82, 2022

  21. [29]

    M. J. Colbrook, V. Antun, and A. C. Hansen. The difficulty of computing stable and accurate neural networks: On the barriers of deep learning and Smale’s 18th problem. Proceedings of the National Academy of Sciences , 119(12):e2107151119, 2022

  22. [30]

    M. J. Colbrook and L. J. Ayton. A contour method for time-fractional PDEs and an application to fractional viscoelastic beam equations. Journal of Computational Physics , 454:110995, 2022

  23. [31]

    M. J. Colbrook and A. C. Hansen. The foundations of spectral computations via the solvability complexity index hierarchy. Journal of the European Mathematical Society, 25(12):4639–4728, 2022

  24. [32]

    M. J. Colbrook, I. Mezi´ c, and A. Stepanenko. Limits and powers of Koopman learning. Preprint: arXiv:2407.06312, 2024

  25. [33]

    M. J. Colbrook, B. Roman, and A. C. Hansen. How to compute spectra with error control. Physical Review Letters, 122(25):250201, 2019

  26. [34]

    M. J. Colbrook and A. Townsend. Avoiding discretization issues for nonlinear eigenvalue problems. SIAM Journal on Matrix Analysis and Applications , 46(1):648–675, 2025

  27. [35]

    Cordes and J

    H. Cordes and J. P. Labrousse. The invariance of the index in the metric space of closed operators. Journal of Mathematics and Mechanics , pages 693–719, 1963

  28. [36]

    Doyle and C

    P. Doyle and C. McMullen. Solving the quintic by iteration. Acta Mathematica, 163(1):151–180, 1989

  29. [37]

    Dyatlov and M

    S. Dyatlov and M. Zworski. Stochastic stability of Pollicott–Ruelle resonances. Nonlinearity, 28(10):3511–3533, 2015

  30. [38]

    L. B. Eldred, W. P. Baker, and A. N. Palazotto. Kelvin-Voigt versus fractional derivative model as constitutive relations for viscoelastic materials. AIAA Journal, 33(3):547–550, 1995

  31. [39]

    Engstr¨ om and M

    C. Engstr¨ om and M. Wang. Complex dispersion relation calculations with the symmetric interior penalty method. Inter. J. Numer. Meth. Eng. , 84(7):849–863, 2010

  32. [40]

    W. F. Fagan, R. S. Cantrell, and C. Cosner. How habitat edges change species interactions. The American Naturalist, 153(2):165–182, 1999

  33. [41]

    T. Faria. Stability and bifurcation for a delayed predator–prey model and the effect of diffusion. Journal of Mathematical Analysis and Applications , 254(2):433–463, 2001

  34. [42]

    Galkowski and M

    J. Galkowski and M. Zworski. Viscosity limits for zeroth-order pseudodifferential operators. Com- munications on Pure and Applied Mathematics , 75(8):1798–1869, 2022

  35. [43]

    S. A. Gershgorin. Uber die Abgrenzung der Eigenwerte einer Matrix. Izvestija Akademii Nauk SSSR, Serija Matematika , 6:749–754, 1931

  36. [44]

    I. T. Gokhberg and M. Krein. The basic propositions on defect numbers, root numbers and indices of linear operators. Uspekhi Mat. Nauk , 12(2):43–118, 1957

  37. [45]

    Guglielmi, M

    N. Guglielmi, M. L´ opez-Fern´ andez, and M. Manucci. Pseudospectral roaming contour integral methods for convection-diffusion equations. Journal of Scientific Computing , 89(1):1–31, 2021. 29

  38. [46]

    A. C. Hansen. On the solvability complexity index, the n-pseudospectrum and approximations of spectra of operators. Journal of the American Mathematical Society , 24(1):81–124, 2011

  39. [47]

    Harari, K

    I. Harari, K. Grosh, T. Hughes, M. Malhotra, P. Pinsky, J. Stewart, and L. Thompson. Recent developments in finite element methods for structural acoustics. Arch. Comput. Meth. Eng. , 3:131– 309, 1996

  40. [48]

    R. Hilfer. Applications of fractional calculus in physics . World scientific, 2000

  41. [49]

    Jarlebring

    E. Jarlebring. The spectrum of delay-differential equations: Numerical methods, stability and per- turbation. PhD thesis, 2008

  42. [50]

    T. Kato. Perturbation theory for nullity, deficiency and other quantities of linear operators. Journal d’Analyse Math´ ematique, 6(1):261–322, 1958

  43. [51]

    T. Kato. Perturbation Theory for Linear Operators, volume 132. Springer Berlin Heidelberg, second edition, 1995

  44. [52]

    M. G. Krein and M. A. Krasnoselski. Fundamental theorems concerning the extension of Hermitian operators and some of their applications to the theory of orthogonal polynomials and the moment problem. Uspekhi Mat. Nauk. , 2:60–106, 1947

  45. [53]

    Kukelova, M

    Z. Kukelova, M. Bujnak, and T. Pajdla. Polynomial eigenvalue solutions to the 5-pt and 6-pt relative pose problems. In BMVC, volume 2, page 2008, 2008

  46. [54]

    Lancaster

    P. Lancaster. Lambda-matrices and vibrating systems. Courier Corporation, 2002

  47. [55]

    Langer, B

    H. Langer, B. Najman, and C. Tretter. Spectral theory of the Klein–Gordon equation in Pontryagin spaces. Communications in mathematical physics , 267(1):159–180, 2006

  48. [56]

    Langer, B

    H. Langer, B. Najman, and C. Tretter. Spectral theory of the Klein–Gordon equation in Krein spaces. Proceedings of the Edinburgh Mathematical Society, 51(3):711–750, 2008

  49. [57]

    Laptev and Y

    A. Laptev and Y. Safarov. Szeg˝ o type limit theorems. Journal of Functional Analysis , 138(2):544– 559, 1996

  50. [58]

    Lewandowski and B

    R. Lewandowski and B. Chorazyczewski. Identification of the parameters of the Kelvin–Voigt and the Maxwell fractional models, used to modeling of viscoelastic dampers. Computers & Structures , 88(1):1–17, 2010

  51. [59]

    Li and F

    C. Li and F. Zeng. Numerical methods for fractional calculus , volume 24. CRC Press, 2015

  52. [60]

    B.-S. Liao, Z. Bai, L.-Q. Lee, and K. Ko. Nonlinear Rayleigh–Ritz iterative method for solving large scale nonlinear eigenvalue problems. Taiwanese J. Math., 14(3A):869–883, 2010

  53. [61]

    Lundberg

    L.-E. Lundberg. Spectral and scattering theory for the Klein-Gordon equation. Communications in Mathematical Physics, 31:243–257, 1973

  54. [62]

    R. L. Magin. Fractional calculus models of complex dynamics in biological tissues. Computers & Mathematics with Applications , 59(5):1586–1593, 2010

  55. [63]

    Mainardi

    F. Mainardi. Fractional calculus and waves in linear viscoelasticity: An introduction to mathematical models. World Scientific, 2010

  56. [64]

    McLean and V

    W. McLean and V. Thom´ ee. Maximum-norm error analysis of a numerical solution via Laplace transformation and quadrature of a fractional-order evolution equation. IMA journal of numerical analysis, 30(1):208–230, 2010

  57. [65]

    McLean and V

    W. McLean and V. Thom´ ee. Numerical solution via Laplace transforms of a fractional order evolu- tion equation. The Journal of Integral Equations and Applications , pages 57–94, 2010

  58. [66]

    McMullen

    C. McMullen. Families of rational maps and iterative root-finding algorithms. Annals of Mathe- matics, 125(3):467–493, 1987

  59. [67]

    McMullen

    C. McMullen. Braiding of the attractor and the failure of iterative algorithms. Inventiones Mathe- maticae, 91(2):259–272, 1988

  60. [68]

    Mehrmann and C

    V. Mehrmann and C. Schr¨ oder. Nonlinear eigenvalue and frequency response problems in industrial practice. J. Math. Ind. , 1:1–18, 2011. 30

  61. [69]

    Metzler and J

    R. Metzler and J. Klafter. The random walk’s guide to anomalous diffusion: A fractional dynamics approach. Physics reports, 339(1):1–77, 2000

  62. [70]

    D. Mugnolo. Abstract wave equations with acoustic boundary conditions. Mathematische Nachrichten, 279(3):299–318, 2006

  63. [71]

    Olver and A

    S. Olver and A. Townsend. A fast and well-conditioned spectral method. 55(3):462–489, 2013

  64. [72]

    Olver and A

    S. Olver and A. Townsend. A practical framework for infinite-dimensional linear algebra. In 2014 First Workshop for High Performance Technical Computing in Dynamic Languages , HPTCDL ’14, pages 57–62, Piscataway, NJ, USA, 2014. IEEE

  65. [73]

    Pang and H.-W

    H.-K. Pang and H.-W. Sun. Fast numerical contour integral method for fractional diffusion equa- tions. J. Sci. Comput. , 66(1):41–66, 2016

  66. [74]

    T. Pritz. Five-parameter fractional derivative model for polymeric damping materials. Journal of Sound and Vibration , 265(5):935–952, 2003

  67. [75]

    R¨ osler and C

    F. R¨ osler and C. Tretter. Computing Klein-Gordon spectra. IMA Journal of Numerical Analysis , 2024

  68. [76]

    J. A. T. M. J. Sabatier, O. P. Agrawal, and J. A. T. Machado. Advances in fractional calculus , volume 4. Springer, 2007

  69. [77]

    Sakoda, N

    K. Sakoda, N. Kawai, T. Ito, A. Chutinan, S. Noda, T. Mitsuyu, and K. Hirao. Photonic bands of metallic systems. I. Principle of calculation and accuracy. Phys. Rev. B , 64(4):045116, 2001

  70. [78]

    Schiff, H

    L. Schiff, H. Snyder, and J. Weinberg. On the existence of stationary states of the mesotron field. Physical Review, 57(4):315, 1940

  71. [79]

    Schwinger

    J. Schwinger. Unitary operator bases. Proceedings of the National Academy of Sciences, 46(4):570– 579, 1960

  72. [80]

    Shargorodsky

    E. Shargorodsky. On the level sets of the resolvent norm of a linear operator. Bulletin of the London Mathematical Society, 40(3):493–504, 2008

  73. [81]

    Sj¨ ostrand and M

    J. Sj¨ ostrand and M. Zworski. Asymptotic distribution of resonances for convex obstacles. Acta Mathematica, 183(2):191–253, 1999

  74. [82]

    S. Smale. The fundamental theorem of algebra and complexity theory. Bulletin of the American Mathematical Society, 4(1):1–36, 1981

  75. [83]

    S. Smale. On the efficiency of algorithms of analysis. Bulletin of the American Mathematical Society, 13(2):87–122, 1985

  76. [84]

    S. Smale. Complexity theory and numerical analysis. Acta Numerica, 6:523–551, 1997

  77. [85]

    Steinbach and G

    O. Steinbach and G. Unger. A boundary element method for the Dirichlet eigenvalue problem of the Laplace operator. 113:281–298, 2009

  78. [86]

    G. Szeg˝ o. Beitr¨ age zur Theorie der Toeplitzschen Formen.Mathematische Zeitschrift, 6(3-4):167– 202, 1920

  79. [87]

    L. N. Trefethen and J. A. C. Weideman. The exponentially convergent trapezoidal rule. SIAM Review, 56(3):385–458, 2014

  80. [88]

    W. Tucker. Validated Numerics. Princeton University Press, 2011

  81. [89]

    A. M. Turing. On computable numbers, with an application to the Entscheidungsproblem. Pro- ceedings of the London Mathematical Society , s2-42(1):230–265, 1937

  82. [90]

    H. Voss. A rational spectral problem in fluid-solid vibration. Preprints des Institutes f¨ ur Mathematik, 2002

  83. [91]

    M. Webb. Isospectral Algorithms, Toeplitz Matrices and Orthogonal Polynomials . PhD thesis, University of Cambridge, 2017

  84. [92]

    Webb and S

    M. Webb and S. Olver. Spectra of Jacobi operators via connection coefficient matrices. 382(2):657– 707, 2021. 31

  85. [93]

    R. Weder. Scattering theory for the Klein-Gordon equation. Journal of Functional Analysis , 27(1):100–117, 1978

  86. [94]

    J. A. C. Weideman. Optimizing Talbot’s contours for the inversion of the Laplace transform. SIAM Journal on Numerical Analysis , 44(6):2342–2362, 2006

  87. [95]

    M. Zworski. Resonances in physics and geometry. Notices of the American Mathematical Society , 46(3):319–328, 1999

  88. [96]

    M. Zworski. Scattering resonances as viscosity limits. In Algebraic and Analytic Microlocal Analysis, pages 635–654. Springer International Publishing, 2018. 32

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.