Pólya’s conjecture for Dirichlet eigenvalues of annuli

Nikolay Filonov, Michael Levitin, Iosif Polterovich, David A. Sher · 2025

Read Pólya’s conjecture for Dirichlet eigenvalues of annuli by Nikolay Filonov, Michael Levitin, Iosif Polterovich, David A. Sher (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-21737.txt.

Introduction

Nikolay Filonov and co-authors prove a classic eigenvalue guess for ring domains in a 2025 paper. Pólya's conjecture says drum tones of a shape never dip below a simple area-based bound. For disks and balls the team had already shown the guess holds.

Annuli, rings between two circles, are the target this time. Variational bounds meet careful estimates of Bessel function phases. Refined lattice point counting and checked computer work close the remaining gaps. A sharper two-term bound for the disk falls out along the way. A sixty-year-old guess gains one more family of shapes.

Catalog record

CREATORNikolay Filonov, Michael Levitin, Iosif Polterovich, David A. Sher
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE72 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-21737

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Pólya’s conjecture for Dirichlet eigenvalues of annuli The accompanying Mathematica script, its printout, and data files are available for download at https://www.michaellevitin.net/polya.html#annuli . MSC(2020): Primary 35P15. Secondary 35P20, 33C10, 11P21, 65G20. Keywords: Laplacian, eigenvalues, Weyl’s law, lattice points, Bessel functions, Bessel phase functions Nikolay Filonov, Michael Levitin, Iosif Polterovich, David A. Sher Abstract We prove Pólya’s conjecture for the eigenvalues of the Dirichlet Laplacian on annular domains. Our approach builds upon and extends the methods we previously developed for disks and balls. It combines variational bounds, estimates of Bessel phase functions, refined lattice point counting techniques, and a rigorous computer-assisted analysis. As a by-product, we also derive a two-term upper bound for the Dirichlet eigenvalue counting function of the disk, improving upon Pólya’s original estimate. \renewpagestyle plain[ ] \footrule \setfoot 1 \newpagestyle mystyle[] \headrule \sethead [1][][Pólya’s conjecture for Dirichlet eigenvalues of annuli]N. Filonov, M. Levitin, I. Polterovich, and D. A. Sher1 §1. Introduction and main results §1.1. Preliminaries Let \Omega\subset\mathbb{R}^{2} be a bounded domain, and let \lambda_{1}(\Omega)\leq\lambda_{2}(\Omega)\leq\dots be the eigenvalues of the Dirichlet Laplacian -\Delta^{\mathrm{Dir}}_{\Omega} in \Omega , enumerated with multiplicities. Denote by \mathcal{N}^{\mathrm{Dir}}_{\Omega}(\lambda):=\#\left\{n\in\mathbb{N}:\lambda_{n}(\Omega)\leq\lambda^{2}\right\} the corresponding eigenvalue counting function. By Weyl’s law, its asymptotics is given by \mathcal{N}^{\mathrm{Dir}}_{\Omega}(\lambda)=\frac{\operatorname{Area}(\Omega)}{4\pi}\lambda^{2}+o\left(\lambda^{2}\right)\qquad\text{as }\lambda\to+\infty. (1.1) More than seventy years ago, George Pólya [ Pó54 ] conjectured that the leading term of the asymptotics ( 1.1 ) is in fact a uniform bound on \mathcal{N}^{\mathrm{Dir}}_{\Omega}(\lambda) for any \Omega\subset\mathbb{R}^{2} : namely, that for all \lambda>0 , \mathcal{N}^{\mathrm{Dir}}_{\Omega}(\lambda)<\frac{\operatorname{Area}(\Omega)}{4\pi}\lambda^{2}. (1.2) Pólya himself later proved this for domains which tile the plane [ Pól61 ] . No other planar domains for which Pólya’s conjecture holds were known until very recently, when the authors of this paper proved it for disks and finite sectors, see [ FLPS23 ] . That paper also gives a more comprehensive account of Pólya’s conjecture for the Dirichlet Laplacian and its counterpart in the Neumann case, as well as its higher-dimensional generalisations. §1.2. Main result The goal of this article is to extend the proof of Pólya’s conjecture to annuli A_{r}=\left\{(x_{1},x_{2})\in\mathbb{R}^{2}:r^{2}<x_{1}^{2}+x_{2}^{2}<1\right\}. with inner radius r\in(0,1) and outer radius one. Note that any annulus is of this form up to rescaling. Note also that r=0 corresponds to the punctured disk which has the same spectrum as the disk (see also [ LMP23 , Example 2.2.23] ) and, therefore, is covered by the results of [ FLPS23 ] . For brevity, we will denote the eigenvalues of an annulus and its eigenvalue counting function by \lambda_{r,n}:=\lambda_{n}(A_{r}),\qquad\mathcal{N}_{r}(\lambda):=\mathcal{N}^{\mathrm{Dir}}_{A_{r}}(\lambda). Given that \operatorname{Area}(A_{r})=\pi(1-r^{2}) , Pólya’s conjecture ( 1.2 ) for annuli states that the inequality \mathcal{N}_{r}(\lambda)<\frac{1-r^{2}}{4}\lambda^{2} (1.3) holds for all pairs (r,\lambda) in the parameter space \mathrm{R}\Lambda:=\left\{(r,\lambda):0<r<1,\lambda>0\right\}. It will be often convenient to work with the parameter \mu:=r\lambda\in[0,\lambda) instead of r . In terms of this parameter, the equivalent form of Pólya’s conjecture for the Dirichlet Laplacian on annuli is that the inequality \mathcal{N}_{\frac{\mu}{\lambda}}(\lambda)<\frac{\lambda^{2}-\mu^{2}}{4} (1.4) holds for all pairs (\lambda,\mu) in the parameter space \Lambda\mathrm{M}:=\left\{(\lambda,\mu):0<\mu<\lambda\right\}. Our main result is Theorem 1.1 . Pólya’s conjecture for the Dirichlet Laplacian holds for all planar annuli. In other words, inequality ( 1.3 ) holds for all (r,\lambda)\in\mathrm{R}\Lambda , or, equivalently, ( 1.4 ) holds for all (\lambda,\mu)\in\Lambda\mathrm{M} . Note that Theorem 1.1 gives a first example of a non-simply connected planar domain for which Pólya’s conjecture has been verified. Remark 1.2 . The eigenvalues of planar annuli have also been investigated in [ GMWW21 ] and more recently in [ GJWY24 ] . Using number-theoretic techniques, these papers improve remainder estimates in the two-term Weyl asymptotics for annuli. Note that the two-term asymptotics implies that Pólya’s conjecture holds for sufficiently large (but unspecified) values of \lambda . While the present paper shares with [ GMWW21 , GJWY24 ] the method of relating the eigenvalue count to a lattice point count, our results are in a way complementary, as we focus on accurate eigenvalue bounds rather than asymptotics. ∎ §1.3. Ideas of the proof The proof of Theorem 1.1 combines analytic arguments with a rigorous computer-assisted verification. We analytically prove Theorem 1.1 in five distinct but partially intersecting regions \mathrm{R}\Lambda_{\aleph} , \aleph=\mathrm{I},\dots,\mathrm{V} , of \mathrm{R}\Lambda . We postpone their exact definitions until later, but show their graphical representation in Figure 1 . Figure 1: Summary of all regions in (r,\lambda) -plane for which Pólya’s conjecture is analytically proven to be true, with two side figures zooming near the axes. The exact definitions of the regions can be found in Theorem 2.1 for \mathrm{R}\Lambda_{\mathrm{I}} , Theorem 2.3 for \mathrm{R}\Lambda_{\mathrm{II}} , Theorem 7.1 for \mathrm{R}\Lambda_{\mathrm{III}} , Theorem 7.2 for \mathrm{R}\Lambda_{\mathrm{IV}} , and Theorem 7.6 for \mathrm{R}\Lambda_{\mathrm{V}} . Different approaches are used to deal with each of these regions. In the region \mathrm{R}\Lambda_{\mathrm{I}} the result follows from the Faber–Krahn and Krahn–Szego isoperimetric inequalities for the first and the second Dirichlet eigenvalues. In the region \mathrm{R}\Lambda_{\mathrm{II}} , we use a comparison between the counting functions for the annuli and for certain carefully chosen flat cylinders, see Lemma 2.2 . We then take advantage of the fact that the eigenvalues of these cylinders can be computed explicitly, see Theorem 2.3 . Note that the arguments in the first two regions are of purely spectral-theoretic nature. The proofs in the remaining three regions are based on results concerning the Bessel phase functions as well as lattice-counting techniques. Here we extend the approach developed in [ FLPS23 ] for the disk. In order to adapt it to the case of annuli we have to overcome several new technical challenges. First, instead of estimating the Bessel phase function itself, we now need to bound the difference between the values of the phase functions at two distinct points, which are chosen depending on the spectral parameter and the inner radius of the annulus (recall that the outer radius is fixed and is equal to one). Some of these bounds follow from the results of [ FLPS24b ] , but some are new, see § 5 . Second, we have to use a more elaborate reduction to a lattice counting problem. In the region \mathrm{R}\Lambda_{\mathrm{III}} we count points below a single curve as in [ FLPS23 ] . In the other two regions, in order to get accurate enough bounds, we need to split the interval [0,\lambda] into two (in \mathrm{R}\Lambda_{\mathrm{IV}} ) and three (in \mathrm{R}\Lambda_{\mathrm{V}} ) subintervals, respectively, and in each subinterval we count integer points under different curves, depending on which bound for the phase functions we are using (see Lemma 5.2 and Figure 6 ). Third, we need to improve upon the lattice counting techniques of [ FLPS23 ] . Recall that [ FLPS23 , Theorem 5.1] provided a way to estimate the number of lattice points under a graph of a non-negative decreasing convex function with a Lipschitz constant \frac{1}{2} . We need to extend this bound to concave functions, as well as to relax the assumption on the Lipschitz constant. We obtain several results in this direction in § 3 . We believe they could be of independent interest, and for illustration we present one of them below. The following definition is central for our analysis. Definition 1.3 . Let a,b\in\mathbb{Z} with a<b , and let g:[a,b]\to\mathbb{R} . The trapezoidal floor sum of g on [a,b] is defined as \mathbf{T}(g,a,b)=\frac{1}{2}\sum_{m=a}^{b-1}\left(\left\lfloor g(m)\right\rfloor+\left\lfloor

Read the complete text