betaFeedback

WorksMathematics

A Kreiss theorem for fractional Cauchy problems with a finite slow component

Abstract

Fractional Cauchy problems with a Caputo derivative of order bounded by one are solved by Mittag-Leffler operator families that lack the semigroup property but remain bounded for classes of sectorial generators, even with spectrum in the right half-plane. To study their transient growth we introduce the fractional Kreiss constant via the Laplace transform of the propagator family. For a closed generator on a Banach space with a time-scale separation and a finite spectral component, that is, a finite-rank projector commuting with the generator whose complement carries a bounded family, we prove that the projected part of the propagator (typically slow spectral part carrying dynamics that is pertinent at mid and long time-scales) is comparable to a deflated Kreiss constant up to a factor linear in the rank, uniformly in fractional order and independently of the dimension. The analytic core of this result is a Bernstein-type inequality for a class of non-rational functions on vertical lines proved by bounding the turning of the derivative. This is a fundamentally different technique from the classical one used in Kreiss theorem for Markov semigroups based on counting intersections of algebraic curves. Importantly, we show that the rank dependence cannot be removed while keeping the Kreiss constant upper bound. Finally, we show that in Hilbert spaces, generators whose numerical range avoids the sector of instability obey the sharp bound of normal operators, which no spectral-set estimate improves. To conclude, we discuss how subdiffusion driven by a self-excited oscillator, the time-fractional Schrödinger equation, and time-changed Markov dynamics illustrate the theory.

Introduction

The stability of a linear evolution equation tells us what happens eventually, while its transient tells us what happens on the way, and for non-normal generators the two can be very far apart. In a standard finite dimensional case, i.e. for a matrix A∈Cd×dA \in\mathbb{C}^{d \times d} the transient growth is bounded by the Kreiss constant K1(A):=sup⁡Re⁡z>0Re⁡(z)∥(z−A)−1∥\mathcal{K}_1(A) := \sup_{\operatorname{Re} z>0} \operatorname{Re}(z) \lVert(z-A)^{-1}\rVert, and the continuous-time Kreiss matrix theorem

K1(A)≤sup⁡t≥0∥etA∥≤ed K1(A)(1)\mathcal{K}_1(A) \le\sup_{t\geq0}\lVert e^{tA}\rVert\le e d\,\mathcal{K}_1(A) \tag*{(1)}

states that the peak of the transient growth is controlled from both sides by a single resolvent quantity, with a factor linear in the dimension [29, 30, 31, 44, 47, 48]. This two-sided estimate is the backbone of the pseudospectral theory of transient growth [47], and its infinite-dimensional limits are by now fully understood: on Banach spaces the Kreiss resolvent condition gives no bound on the semigroup at all, while on Hilbert spaces it gives sublinear growth in time [1, 2, 9, 13].

The purpose of this paper is to develop the corresponding theory for fractional Cauchy problems

∂tαu=Lu,u(0)=u0,0<α≤1,(2)\partial_t^\alpha u = Lu,\qquad u(0)=u_0,\qquad0<\alpha\leq1, \tag*{(2)}

with the Caputo derivative ∂tα\partial_t^\alpha and a closed operator LL on a Banach space XX. Such problems describe evolutions with memory: subdiffusive transport in heterogeneous or crowded media arises as the scaling limit of continuous-time random walks with heavy-tailed waiting times [33, 34, 36], the time-fractional Schrödinger equation was proposed as a quantum model with memory [37], and fractional relaxation and oscillation equations are standard models in viscoelasticity and control [17, 32, 38]. The solution of (2) is u(t)=Eα(tαL)u0u(t)=E_\alpha(t^\alpha L)u_0, where Eα(tαL)E_\alpha(t^\alpha L) is the Mittag–Leffler family generated by LL [4, 40], given for bounded or sectorial LL by the Mittag–Leffler function Eα(z)=∑k≥0zkΓ(1+αk)E_\alpha(z)=\sum_{k\geq0}\frac{z^k}{\Gamma(1+\alpha k)}.

Already at the level of definitions, these families differ from semigroups in ways that call for a theory of their own. The map t↦Eα(tαL)t\mapsto E_\alpha(t^\alpha L) has no semigroup property, so arguments that propagate a bound from short to long times are not available. The stability region is larger: for a matrix LL the family is bounded if and only if the spectrum avoids the open sector Σα:={z≠0:∣Arg⁡z∣<απ2}\Sigma_\alpha:=\{z\ne0:|\operatorname{Arg}z|<\frac{\alpha\pi}{2}\} and the eigenvalues on its boundary are semisimple [32]. Since Σα\Sigma_\alpha is a proper subsector of the right half-plane when α<1\alpha<1, modes that grow exponentially in physical time can be perfectly stable under the fractional clock, and the subordination formula, which writes Eα(tαL)E_\alpha(t^\alpha L) as an average of esLe^{sL} over a random time [4], then gives no information, because it averages an exponentially growing semigroup.

The envelope t↦∥Eα(tαL)∥t\mapsto\lVert E_\alpha(t^\alpha L)\rVert in general has a short-time, an intermediate and a long-time regime, each governed by a different resolvent quantity. In particular, in Hilbert spaces for the semigroup case α=1\alpha=1, it is well understood that numerical and spectral abscissa govern short-time and long-time behavior, respectively, see e.g. [47]. While this remains also the case in the fractional setting (we review these essentially known results in Section Appendix B.3), the characterization of intermediate behavior is much more involved problem. Indeed, whenever the operator LL is normal, while the envelope of the semigroup ∥etL∥\lVert e^{tL}\rVert is sharply bounded by one, in the fractional setting ∣Eα∣|E_\alpha| exceeds 1/α1/\alpha at points of the boundary rays of Σα\Sigma_\alpha, implying that even a normal generator with spectrum on these rays amplifies envelope by more than 1/α1/\alpha, thus provoking transient growth. Hence, in this paper we focus on answering the question:

Which resolvent quantity controls sup⁡t≥0∥Eα(tαL)∥\sup_{t\ge0}\|E_\alpha(t^\alpha L)\|, and with which constants?

In answering it, we particularly focus on the finite-rank operators on infinite-dimensional spaces, which is motivated by recent works on data-driven estimators of transfer operators and generators of dynamical systems [21, 23, 24, 26], including estimators of slow modes of molecular systems [11], and their pseudospectra can be computed at a cost governed by the rank [27].

Related work

The continuous-time Kreiss matrix theorem goes back to Kreiss [29, 30]; the linear dependence on the dimension was obtained by LeVeque and Trefethen [31], the constant ee through Spijker’s [44, 48, 50], and Kraaijevanger [28] showed that the linear order is sharp, for powers as well as for the exponential; see also Spijker et al. [43], and for general domains Toh and Trefethen [46]. Eisner and Zwart [13] showed that the Kreiss condition gives no growth bound on Banach spaces, that it gives ∥etL∥=O(t)\|e^{tL}\|=O(t) on Hilbert spaces, and that growth of order tγt^\gamma occurs for every γ<1\gamma<1, using Fourier multipliers on L2L^2 spaces with power weights of class A2A_2. The Hilbert-space bound was improved to O(tlog⁡t)O\left(\frac{t}{\sqrt{\log t}}\right) by Arnold [1], extended to general resolvent growth by Dell’Oro [9] and Dell’Oro et al. [10], and very recently to O(t1−ε)O(t^{1-\varepsilon}) with ε\varepsilon depending on the Kreiss constant [2]; see also Rozendaal and Veraar [41]. All these arguments use the semigroup law. On the side of spectral sets, the numerical range is a (1+2)(1+\sqrt{2})-spectral set [8], von Neumann’s inequality makes a half-plane containing the numerical range a spectral set with constant one [45, 49], and Crouzeix [7] interpolates between the two for unbounded convex sets.

Mittag–Leffler families, also called α\alpha-times resolvent families or solution operators, are the resolvent families of Volterra equations with the kernel tα−1Γ(α)\frac{t^{\alpha-1}}{\Gamma(\alpha)} [4, 40]. Their generation theory, subordination and asymptotics are classical [4, 17, 40], and resolvent characterisations of growth bounds and of uniform stability were obtained recently by Ghavidel and Shahali [16] and Mei and Li [35], in the spirit of the Hille–Yosida and Gearhart–Prüss theorems. These results are qualitative or asymptotic, and we are not aware of quantitative two-sided bounds of Kreiss type for Mittag–Leffler families beyond the matrix case conjectured in Kostić and Uzelac Bujišić [22]. For matrices, the spectral criterion of Matignon [32] characterises boundedness, and fractional pseudospectra have been used to localise spectra and to study stability of fractional systems [12, 42].

Organisation

For the reader’s convince, we start with the summerize of our main results as four Theorems A–D in § 2. Then, in § 3 we review facts on the Mittag–Leffler functions, Mittag–Leffler operator families, and prove basic properties of the newly introduced the fractional Kreiss constant. Based on these preliminary results, § 4 proves Theorem B, and § 5 the finite-rank core of Theorem A. § 6 treats normal operators and the numerical range, and § 8 completes Theorem A for generators with a slow component and analyses prototype (a). § 9 proves Theorem C, § 10 describes the mechanisms and time scales of the transient behavior, § 11 revisits the three prototypes introduced in § 2, and § 12 discusses open problems. The appendix contains the table of notation and the proofs of auxiliary results, and the Supplementary Material the numerical methods, the complete tables, the study of random finite-rank operators and a description of the code.

Main contributions in a nutshell

Recalling that the Laplace transform of Eα(tαL)E_\alpha(t^\alpha L) is ωα−1(ωα−L)−1\omega^{\alpha-1}(\omega^\alpha-L)^{-1}, somewhat natural way to define the fractional Kreiss constant is

Kα(L):=sup⁡Re⁡ω>0Re⁡(ω)∥ωα−1(ωα−L)−1∥=sup⁡z∈Σαaα(z)∥(z−L)−1∥,(3)\mathcal{K}_\alpha(L) := \sup_{\operatorname{Re}\omega>0} \operatorname{Re}(\omega)\left\|\omega^{\alpha-1}(\omega^\alpha-L)^{-1}\right\| = \sup_{z\in\Sigma_\alpha} a_\alpha(z)\left\|(z-L)^{-1}\right\|, \tag*{(3)}

where aα(z):=∣z∣cos⁡(Arg⁡zα)a_\alpha(z):=|z|\cos\left(\frac{\operatorname{Arg}z}{\alpha}\right); it reduces to K1\mathcal{K}_1 for α=1\alpha=1. The second expression replaces the classical weight Re⁡z\operatorname{Re}z on the right half-plane by a weight on Σα\Sigma_\alpha that vanishes on the boundary rays, so Kα(L)\mathcal{K}_\alpha(L) is finite for operators that are unstable as generators of semigroups. Up to a factor between 11 and 1α\frac{1}{\alpha}, aα(z)a_\alpha(z) is the distance from zz to the boundary of Σα\Sigma_\alpha (Theorem 3.6(c)), so Kα(L)\mathcal{K}_\alpha(L) measures how fast the resolvent may blow up as zz approaches the stability region from inside the sector of instability. A constant of this type, for matrices and in pseudospectral form, was proposed in the precursor [22], where a fractional Kreiss matrix theorem was conjectured; the present paper proves this conjecture in a stronger form, for all real α∈(0,1]\alpha\in(0,1], on Banach spaces and with the rank in place of the dimension.

Our main object is a generator with a finite slow component, i.e. a system whose mid and long-term dynamics is determined by finitely many eigenvalues that are well separated from the fast time-scales. More precisely, a projection P∈B(X)P\in\mathcal{B}(X) of finite rank mm is a slow component of LL if it commutes with LL, the part of LL in (I−P)X(I-P)X generates a bounded Mittag–Leffler family, spectra of PLPL and (I−P)L(I-P)L are disjoint with the former being closer to stability boundary.

As we first show the transient growth of such a generator is carried out by the mm slow modes and by the geometry of the splitting, and both are seen by the deflated Kreiss constant

Kα(L;P):=sup⁡z∈Σαaα(z)∥(z−L)−1P∥,\mathcal{K}_\alpha(L;P):=\sup_{z\in\Sigma_\alpha}a_\alpha(z)\left\|(z-L)^{-1}P\right\|,

the Kreiss quantity of the compressed resolvent.

Theorem A (Theorems 5.1, 5.4 and 7.1). Let 0<α≤10<\alpha\leq1, let X≠{0}X\neq\{0\} be a Banach space, let LL be a closed, densely defined operator on XX, and put

κα,n:=e2(3+4n(2−α)),κα,n′:=e2(1+4n(2−α)).\kappa_{\alpha,n}:=\frac{\mathrm{e}}{2}\left(3+4n(2-\alpha)\right),\qquad\kappa'_{\alpha,n}:=\frac{\mathrm{e}}{2}\left(1+4n(2-\alpha)\right).

(i) If LL generates a bounded Mittag–Leffler family, then Kα(L)≤sup⁡t≥0∥Eα(tαL)∥\mathcal{K}_\alpha(L)\leq\sup_{t\geq0}\left\|E_\alpha(t^\alpha L)\right\|.

(ii) If PP is a slow component of rank mm, then LL generates a Mittag–Leffler family,

Kα(L;P)≤sup⁡t≥0∥Eα(tαL)P∥≤κα,mKα(L;P),\mathcal{K}_\alpha(L;P)\leq\sup_{t\geq0}\left\|E_\alpha(t^\alpha L)P\right\|\leq\kappa_{\alpha,m}\mathcal{K}_\alpha(L;P),
sup⁡t≥0∥Eα(tαL)∥≤κα,mKα(L;P)+sup⁡t≥0∥Eα(tαL)(I−P)∥,\sup_{t\geq0}\left\|E_\alpha(t^\alpha L)\right\|\leq\kappa_{\alpha,m}\mathcal{K}_\alpha(L;P)+\sup_{t\geq0}\left\|E_\alpha(t^\alpha L)(I-P)\right\|,

and κα,m\kappa_{\alpha,m} may be replaced by κα,m′\kappa'_{\alpha,m} when 00 is not an eigenvalue of the slow block.

(iii) If LL is bounded with rank⁡L≤n\operatorname{rank} L \le n, then sup⁡t≥0∥Eα(tαL)∥≤κα,nKα(L)\sup_{t\ge0}\|E_{\alpha}(t^{\alpha}L)\| \le\kappa_{\alpha,n}\mathcal{K}_{\alpha}(L).

The constants depend on α\alpha and on the rank, never on the dimension of XX, and they are uniform in α∈(0,1]\alpha\in(0,1].

Part (iii) is the core case of (ii): a finite-rank operator with finite Kreiss constant splits as X=Ran⁡L⊕ker⁡LX = \operatorname{Ran} L \oplus\ker L, its slow component is the projection onto Ran⁡L\operatorname{Ran} L along ker⁡L\ker L, and the stationary contribution of the kernel is absorbed into the estimate without paying the norm of the oblique complementary projection. Part (ii) applies to unbounded generators that may have infinite rank, need not generate a C0C_0-semigroup and may have spectrum in the right half-plane, as in prototype (a), and the fast part may be controlled on Banach spaces by sectoriality or on Hilbert spaces by the numerical range.

References

  1. [1]Loris Arnold. Behavior of Kreiss bounded C₀-semigroups on a Hilbert space. Adv. Oper. Theory, 7:Paper No. 62, 2022. doi: 10.1007/s43036-022-00223-z.DOI
  2. [2]Loris Arnold. Polynomial gaps below linear growth for Kreiss bounded semigroups and operators. arXiv:2608.13397, 2026.arxiv.org/abs/2608.13397
  3. [3]Stefan Banach. Sur les lignes rectifiables et les surfaces dont l’aire est finie. Fund. Math., 7:225–236, 1925. doi: 10.4064/fm-7-1-225-236.DOI
  4. [4]Emilia G. Bazhlekova. Fractional Evolution Equations in Banach Spaces. PhD thesis, Technische Universiteit Eindhoven, 2001.DOI
  5. [5]N. H. Bingham. Limit theorems for occupation times of Markov processes. Z. Wahrscheinlichkeitstheorie verw. Gebiete, 17(1):1–22, 1971. doi: 10.1007/BF00538470.DOI
  6. [6]John B. Conway. Functions of One Complex Variable I, volume 11 of Graduate Texts in Mathematics. Springer, New York, second edition, 1978. doi: 10.1007/978-1-4612-6313-5.DOI
  7. [7]Michel Crouzeix. Numerical ranges and spectral sets: the unbounded case. arXiv:2509.19792, 2025.arxiv.org/abs/2509.19792
  8. [8]Michel Crouzeix and César Palencia. The numerical range is a (1 + √2)-spectral set. SIAM J. Matrix Anal. Appl., 38(2):649–655, 2017. doi: 10.1137/17M1116672.DOI
  9. [9]Filippo Dell’Oro. A remark on the growth rate of operator semigroups under resolvent bounds. Semigroup Forum, 111(1):78–84, 2025. doi: 10.1007/s00233-025-10549-8. arXiv:2505.12807.DOI
  10. [10]Filippo Dell’Oro, Alberto Farina, and Vittorino Pata. Unified growth rates for operator semigroups under generalized Kreiss conditions. arXiv:2608.16217, 2026.arxiv.org/abs/2608.16217
  11. [11]Timothée Devergne, Vladimir Kostić, Massimiliano Pontil, and Michele Parrinello. Slow dynamical modes from static averages. J. Chem. Phys., 162(12):Paper No. 124108, 2025. doi: 10.1063/5.0246248.DOI
  12. [12]Kai Diethelm, Safoura Hashemishahraki, Ha Duc Thai, and Hoang The Tuan. A constructive approach for investigating the stability of incommensurate fractional differential systems. J. Math. Anal. Appl., 540:Paper No. 128642, 2024. doi: 10.1016/j.jmaa.2024.128642.DOI
  13. [13]Tanja Eisner and Hans Zwart. Continuous-time Kreiss resolvent condition on infinite-dimensional spaces. Math. Comp., 75(256):1971–1985, 2006. doi: 10.1090/S0025-5718-06-01862-X.DOI
  14. [14]Klaus-Jochen Engel and Rainer Nagel. One-Parameter Semigroups for Linear Evolution Equations, volume 194 of Graduate Texts in Mathematics. Springer, New York, 2000. doi: 10.1007/b97696.DOI
  15. [15]William Feller. An Introduction to Probability Theory and Its Applications, volume II. Wiley, New York, second edition, 1971.
  16. [16]Seyedeh Marzieh Ghavidel and Mina Shahali. Growth bounds for α-times resolvent families. Semigroup Forum, 107:109–126, 2023. doi: 10.1007/s00233-023-10372-z.DOI
  17. [17]Rudolf Gorenflo, Anatoly A. Kilbas, Francesco Mainardi, and Sergei Rogosin. Mittag-Leffler Functions, Related Topics and Applications. Springer Monographs in Mathematics. Springer, Berlin, Heidelberg, second edition, 2020. doi: 10.1007/978-3-662-61550-8.
  18. [18]Loukas Grafakos. Classical Fourier Analysis, volume 249 of Graduate Texts in Mathematics. Springer, New York, third edition, 2014. doi: 10.1007/978-1-4939-1194-3.DOI
  19. [19]Richard Hunt, Benjamin Muckenhoupt, and Richard Wheeden. Weighted norm inequalities for the conjugate function and Hilbert transform. Trans. Amer. Math. Soc., 176:227–251, 1973. doi: 10.1090/S0002-9947-1973-0312139-8.
  20. [20]Tosio Kato. Perturbation Theory for Linear Operators. Classics in Mathematics. Springer, Berlin, 1995. doi: 10.1007/978-3-642-66282-9. Reprint of the 1980 edition.DOI
  21. [21]Stefan Klus, Feliks Nüske, Sebastian Peitz, Jan-Hendrik Niemann, Cecilia Clementi, and Christof Schütte. Data-driven approximation of the Koopman generator: Model reduction, system identification, and control. Phys. D, 406:Paper No. 132416, 2020. doi: 10.1016/j.physd.2020.132416.DOI
  22. [22]Vladimir Kostić and Ilja Uzelac Bujišić. Kreissova konstanta za Mittag-Lefflerove funkcije. In META 2025: The 10th Conference on Mathematics in Engineering: Theory and Applications, Conference Proceedings (Novi Sad, Serbia, December 12–14, 2025), pages 108–113, Novi Sad, 2025. Faculty of Technical Sciences, University of Novi Sad. doi: 10.24867/META.2025.18. In Serbian.DOI
  23. [23]Vladimir Kostić, Pietro Novelli, Andreas Maurer, Carlo Ciliberto, Lorenzo Rosasco, and Massimiliano Pontil. Learning dynamical systems via Koopman operator regression in reproducing kernel Hilbert spaces. In Advances in Neural Information Processing Systems, volume 35, pages 4017–4031, 2022.
  24. [24]Vladimir Kostić, Karim Lounici, Pietro Novelli, and Massimiliano Pontil. Sharp spectral rates for Koopman operator learning. In Advances in Neural Information Processing Systems, volume 36, pages 32328–32339, 2023.arxiv.org/abs/2302.02004
  25. [25]Vladimir R. Kostić, Karim Lounici, Hélène Halconruy, Timothée Devergne, and Massimiliano Pontil. Learning the infinitesimal generator of stochastic diffusion processes. In Advances in Neural Information Processing Systems, volume 37, pages 137806–137846, 2024.arxiv.org/abs/2405.12940
  26. [26]Vladimir R. Kostić, Karim Lounici, Hélène Halconruy, Timothée Devergne, Pietro Novelli, and Massimiliano Pontil. Laplace transform based low-complexity learning of continuous Markov semigroups. In Proceedings of the 42nd International Conference on Machine Learning, volume 267 of Proceedings of Machine Learning Research, pages 31560–31589. PMLR, 2025.arxiv.org/abs/2410.14477
  27. [27]Vladimir R. Kostić, Dragana Lj. Cvetković, and Ljiljana Cvetković. Scalable pseudospectral analysis via low-rank approximations of dynamical systems. arXiv:2602.01721, 2026.arxiv.org/abs/2602.01721
  28. [28]J. F. B. M. Kraaijevanger. Two counterexamples related to the Kreiss matrix theorem. BIT, 34(1):113–119, 1994. doi: 10.1007/BF01935020.DOI
  29. [29]Heinz-Otto Kreiss. Über Matrizen die beschränkte Halbgruppen erzeugen. Math. Scand., 7:71–80, 1959.DOI
  30. [30]Heinz-Otto Kreiss. Über die Stabilitätsdefinition für Differenzengleichungen die partielle Differentialgleichungen approximieren. BIT, 2(3):153–181, 1962. doi: 10.1007/BF01957330.DOI
  31. [31]Randall J. LeVeque and Lloyd N. Trefethen. On the resolvent condition in the Kreiss matrix theorem. BIT, 24(4):584–591, 1984. doi: 10.1007/BF01934916.DOI
  32. [32]Denis Matignon. Stability results for fractional differential equations with applications to control processing. In Computational Engineering in Systems Applications (CESA ’96), IMACS/IEEE-SMC Multiconference, Lille, France, July 1996, volume 2, pages 963–968, 1996.
  33. [33]Mark M. Meerschaert and Hans-Peter Scheffler. Limit theorems for continuous-time random walks with infinite mean waiting times. J. Appl. Probab., 41(3):623–638, 2004. doi: 10.1239/jap/1091543414.DOI
  34. [34]Mark M. Meerschaert and Alla Sikorskii. Stochastic Models for Fractional Calculus, volume 43 of De Gruyter Studies in Mathematics. De Gruyter, Berlin, second edition, 2019. doi: 10.1515/9783110560244. First edition 2012.DOI
  35. [35]Jie Mei and Miao Li. Uniform stability and decay rate of solutions for fractional Cauchy problems. Banach J. Math. Anal., 18:Paper No. 50, 2024. doi: 10.1007/s43037-024-00355-9.DOI
  36. [36]Ralf Metzler and Joseph Klafter. The random walk’s guide to anomalous diffusion: a fractional dynamics approach. Phys. Rep., 339(1):1–77, 2000. doi: 10.1016/S0370-1573(00)00070-3.DOI
  37. [37]Mark Naber. Time fractional Schrödinger equation. J. Math. Phys., 45(8):3339–3352, 2004. arXiv:math-ph/0410028.DOI
  38. [38]Igor Podlubny. Fractional Differential Equations, volume 198 of Mathematics in Science and Engineering. Academic Press, San Diego, 1999.
  39. [39]Harry Pollard. The completely monotonic character of the Mittag-Leffler function Eₐ(−x). Bull. Amer. Math. Soc., 54(12):1115–1116, 1948.DOI
  40. [40]Jan Prüss. Evolutionary Integral Equations and Applications, volume 87 of Monographs in Mathematics. Birkhäuser, Basel, 1993. doi: 10.1007/978-3-0348-8570-6.DOI
  41. [41]Jan Rozendaal and Mark Veraar. Sharp growth rates for semigroups using resolvent bounds. J. Evol. Equ., 18:1721–1744, 2018. doi: 10.1007/s00028-018-0459-x.DOI
  42. [42]Ernest Šanca, Vladimir R. Kostić, and Ljiljana Cvetković. Fractional pseudospectra and their localizations. Linear Algebra Appl., 559:244–269, 2018. doi: 10.1016/j.laa.2018.08.035.DOI
  43. [43]M. N. Spijker, S. Tracogna, and B. D. Welfert. About the sharpness of the stability estimates in the Kreiss matrix theorem. Math. Comp., 72(242):697–713, 2003. doi: 10.1090/S0025-5718-02-01472-2.DOI
  44. [44]Marc N. Spijker. On a conjecture by LeVeque and Trefethen related to the Kreiss matrix theorem. BIT, 31(3):551–555, 1991. doi: 10.1007/BF01933268.DOI
  45. [45]Béla Sz.-Nagy, Ciprian Foias, Hari Bercovici, and László Kérchy. Harmonic Analysis of Operators on Hilbert Space. Universitext. Springer, New York, second edition, 2010. doi: 10.1007/978-1-4419-6094-8.
  46. [46]Kim-Chuan Toh and Lloyd N. Trefethen. The Kreiss matrix theorem on a general complex domain. SIAM J. Matrix Anal. Appl., 21(1):145–165, 1999. doi: 10.1137/S0895479897324020.DOI
  47. [47]Lloyd N. Trefethen and Mark Embree. Spectra and Pseudospectra: The Behavior of Nonnormal Matrices and Operators. Princeton University Press, Princeton, NJ, 2005.
  48. [48]J. L. M. van Dorsselaer, J. F. B. M. Kraaijevanger, and M. N. Spijker. Linear stability analysis in the numerical solution of initial value problems. Acta Numer., 2:199–237, 1993. doi: 10.1017/S0962492900002361.DOI
  49. [49]John von Neumann. Eine Spektraltheorie für allgemeine Operatoren eines unitären Raumes. Math. Nachr., 4:258–281, 1951.
  50. [50]Elias Wegert and Lloyd N. Trefethen. From the Buffon needle problem to the Kreiss matrix theorem. Amer. Math. Monthly, 101(2):132–139, 1994. doi: 10.2307/2324361.DOI

Paper details

vlnc:W_yy49k57e
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Claude Opus 5.5, Claude Code

View statement
Looking for related papers…
Contents