Improved lower bounds for Dirichlet eigenvalues of the Laplacian and poly-Laplacian on bounded Euclidean domains

Zhengchao Ji, Yong Luo · 2026

Read Improved lower bounds for Dirichlet eigenvalues of the Laplacian and poly-Laplacian on bounded Euclidean domains by Zhengchao Ji, Yong Luo (2026) 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-2604-20367.txt.

Introduction

Ji and Luo prove improved lower bounds of Brezin-Li-Yau type for mean sums of Dirichlet eigenvalues for the Laplacian and poly-Laplacian on bounded Euclidean domains. The key step is expanding two binary polynomials that may interest readers on their own. Their bounds improve several existing bounds of this kind in the literature. They are optimal in the sense that the expansions capture all positive terms, whereas earlier work captured only a subset of the positive terms in these two polynomials.

Catalog record

CREATORZhengchao Ji, Yong Luo
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE50685
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.20367

Browse the full library · About the Tower of Babel

Full text

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

Improved lower bounds for Dirichlet eigenvalues of the Laplacian and poly-Laplacian on bounded Euclidean domains Zhengchao Ji, Yong Luo Abstract In this paper, we establish Brezin-Li-Yau type lower bounds for averaged sums of Dirichlet eigenvalues of the Laplacian and poly-Laplacian on bounded domains in Euclidean spaces. By deriving expansions of two binary polynomials which may be of independent interest, we improve several existing lower bounds of this kind in the literature. Furthermore, our lower bounds are optimal in the sense that our expansions capture all positive terms, whereas previous works only provided certain lower bounds for these two binary polynomials, effectively capturing only a subset of the positive terms identified in our expansions. MSC 2010 subject classification: 35P15, 58G05 Keywords: The Laplacian, the poly-Laplacian, eigenvalues, Pólya’s conjecture 1 Introduction Let \Omega be a bounded domain with piecewise smooth boundary \partial\Omega in an n -dimensional Euclidean space \mathbb{R}^{n} . First of all, we focus on the following Dirichlet eigenvalue problem of the Laplacian (1.1) \begin{cases}\Delta u=-\lambda u\ &\mathrm{in}\ \Omega,\\ u=0\ &\mathrm{on}\ \partial\Omega.\end{cases} It is well known that spectrum of the eigenvalue problem ( 1.1 ) is real and discrete (cf. [ 4 , 17 , 31 ] ) \displaystyle 0<\lambda_{1}<\lambda_{2}\leq\lambda_{3}\leq\cdots\rightarrow\infty, where each eigenvalue is repeated according to its multiplicity. Let V(\Omega) (or V ) be the volume of \Omega , and \omega_{n} the volume of the unit ball in \mathbb{R}^{n} . Then the following well-known Weyl’s asymptotic formula holds \displaystyle\lambda_{k}\sim\frac{4\pi^{2}}{(\omega_{n}V(\Omega))^{\frac{2}{n}}}k^{\frac{2}{n}},\ k\rightarrow\infty, which implies that (1.2) \displaystyle\frac{1}{k}\sum_{i=1}^{k}\lambda_{i}\sim\frac{n}{n+2}\frac{4\pi^{2}}{(\omega_{n}V(\Omega))^{\frac{2}{n}}}k^{\frac{2}{n}},\,\,\,\,\,k\rightarrow\infty. In 1961, Pólya [ 33 ] proved that, if n=2 and \Omega is a tiling domain in \mathbb{R}^{2} , then \displaystyle\lambda_{k}\geq\frac{4\pi^{2}}{(\omega_{n}V(\Omega))^{\frac{2}{n}}}k^{\frac{2}{n}},\ \ \mathrm{for}\ k=1,2,\ldots. Based on the result above, he proposed the famous conjecture: Pólya’s conjecture If \Omega is a bounded domain in \mathbb{R}^{n} , then the k -th eigenvalue \lambda_{k} of the eigenvalue problem ( 1.1 ) satisfies \displaystyle\lambda_{k}\geq\frac{4\pi^{2}}{(\omega_{n}V(\Omega))^{\frac{2}{n}}}k^{\frac{2}{n}},\ \ \mathrm{for}\ k=1,2,\ldots. During the past six decades, many mathematicians have focused on this problem and related topics, and there are a lot of important results on this subject (cf. [ 2 , 3 , 7 , 8 , 9 , 12 , 13 , 24 , 25 , 27 , 28 ] ). We suggest that readers refer [ 36 , 39 ] for more details. In 1983, Li and Yau [ 26 ] verified the famous inequality (1.3) \displaystyle\frac{1}{k}\sum_{i=1}^{k}\lambda_{i}\geq\frac{n}{n+2}\frac{4\pi^{2}}{(\omega_{n}V(\Omega))^{\frac{2}{n}}}k^{\frac{2}{n}},\ k=1,2,\ldots. By using Legendre transformation of an earlie result obtained by Berezin [ 2 ] , one can also obtain the above inequality. Nowadays the above inequality is called the Brezin-Li-Yau’s inequality in the literature. From ( 1.3 ), one can derive that \displaystyle\lambda_{k}\geq\frac{n}{n+2}\frac{4\pi^{2}}{(\omega_{n}V(\Omega))^{\frac{2}{n}}}k^{\frac{2}{n}},\ \ \mathrm{for}\ k=1,2,\ldots, which gives a partial solution to the Pólya’s conjecture with a factor \frac{n}{n+2} . Recently, Filonov, Levitin, Polterovich and Sher [ 10 , 11 ] made some breakthroughs by solving the Pólya’s Conjecture for balls in \mathbb{R}^{n} and annuli in \mathbb{R}^{2} . In [ 29 ] , Melas obtained the following beautiful estimate which improves ( 1.3 ) for n\geq 2 and k\geq 1 (1.4) \displaystyle\frac{1}{k}\sum_{i=1}^{k}\lambda_{i}\geq\frac{n}{n+2}\frac{4\pi^{2}}{(\omega_{n}V(\Omega))^{\frac{2}{n}}}k^{\frac{2}{n}}+c_{n}\frac{V(\Omega)}{I(\Omega)},\ \ \mathrm{for}\ k=1,2,\ldots, where c_{n}\leq 1/(24(n+2)) is a positive constant depending only on n and \displaystyle I(\Omega)=\min_{a\in\mathbb{R}^{n}}\int_{\Omega}|x-a|^{2}dx is called the moment of inertia of \Omega . Due to [ 14 , 15 , 30 , 35 ] , under suitable conditions on \Omega , we have the following asymptotic formula \displaystyle\frac{1}{k}\sum_{i=1}^{k}\lambda_{i}\sim\frac{n}{n+2}\frac{4\pi^{2}k^{\frac{2}{n}}}{(\omega_{n}V(\Omega))^{\frac{2}{n}}}+\frac{\sqrt{\pi}\Gamma\left(\frac{4+n}{2}\right)^{1+\frac{1}{n}}|\partial\Omega|k^{\frac{1}{n}}}{(n+1)\Gamma\left(\frac{3+n}{2}\right)\Gamma(2)^{\frac{1}{n}}V(\Omega)^{1+\frac{1}{n}}},\,\,\,\,k\rightarrow\infty. Hence, it is natural to find the following type lower bound \displaystyle\frac{1}{k}\sum_{i=1}^{k}\lambda_{i}\geq\frac{n}{n+2}\frac{4\pi^{2}k^{\frac{2}{n}}}{(\omega_{n}V(\Omega))^{\frac{2}{n}}}+\frac{\sqrt{\pi}\Gamma\left(\frac{4+n}{2}\right)^{1+\frac{1}{n}}|\partial\Omega|k^{\frac{1}{n}}}{(n+1)\Gamma\left(\frac{3+n}{2}\right)\Gamma(2)^{\frac{1}{n}}V(\Omega)^{1+\frac{1}{n}}},\,\,\,\,k=1,2,\cdots. In 2009, Kovařík, Vugalter and Weidl [ 24 ] first obtained this kind of results on polygons and established that (1.5) \displaystyle\frac{1}{k}\sum_{i=1}^{k}\lambda_{i}\geq\frac{2\pi}{V(\Omega)}k+4\bar{c}k^{\frac{1}{2}-\varepsilon(k)}V(\Omega)^{-\frac{3}{2}}\sum^{m}_{j=1}l_{j}\Theta\left(k-\frac{9V}{2\pi d^{2}_{j}}\right), where m,l_{j},d_{j},\Theta(x) are some constants depending on \partial\Omega and \displaystyle\varepsilon(k)=\frac{2}{\sqrt{\log_{2}(\frac{2\pi k}{c})}},\,\,\,c=\sqrt{\frac{3\pi}{14}}10^{-11},\,\,\,\bar{c}=\frac{2^{-3}}{9\sqrt{2}\cdot 36}(2\pi)^{\frac{5}{4}}c^{\frac{1}{4}}. Compared with the second term on the right hand of ( 1.4 ) with that of ( 1.5 ), Kovařík-Vugalter-Weidl’s lower bound is sharper for large k . In 2010, Ilyin [ 16 ] improved Melas’s lower bound when n=2,3,4 by proving that (1.6) \displaystyle\frac{1}{k}\sum_{i=1}^{k}\lambda_{i}\geq\frac{n}{n+2}\frac{4\pi^{2}}{(\omega_{n}V(\Omega))^{\frac{2}{n}}}k^{\frac{2}{n}}+\beta_{n}^{L}\frac{n}{48}\frac{V(\Omega)}{I(\Omega)}, where \beta_{n}^{L}=119/120,\beta_{n}^{L}=0.986,\beta_{n}^{L}=0.978 for n=2,3,4 respectively. In 2013, Yildirim and Yolcu [ 41 ] improved ( 1.4 ) by adding the following term to the right hand of ( 1.4 ) (1.7) \displaystyle\begin{split}\frac{V(\Omega)^{\frac{3n+2}{2n}}}{144(n+2)I^{\frac{3}{2}}\Gamma^{\frac{1}{n}}(1+n/2)}k^{-\frac{1}{n}}.\end{split} In 2020, the first author and Xu [ 19 ] proved a sharper polynomial inequality and improved ( 1.7 ) for any bounded domain \Omega\subseteq\mathbb{R}^{n} , n\geq m+1\geq 3 and k\geq 1 as (1.8) \displaystyle\begin{split}\frac{1}{k}\sum_{i=1}^{k}\lambda_{i}\geq&\omega_{n}^{-\frac{2}{n}}{\alpha^{-\frac{2}{n}}}k^{\frac{2}{n}}-\frac{2\omega_{n}^{\frac{m-1}{n}}S_{m+2}\alpha^{\frac{(m+1)n+m-1}{n}}}{(n+2)\rho^{m+1}}k^{\frac{-m+1}{n}}\\ &+c_{2}\frac{2\omega_{n}^{\frac{m}{n}}(m+1)S_{m+3}\alpha^{\frac{(m+2)n+m}{n}}}{(n+2)(m+3)\rho^{m+2}}k^{\frac{-m}{n}},\end{split} where S_{l}=(a+1)^{l}-a^{l}, \displaystyle\begin{split}c_{2}\leq&\min\left\{1,\frac{(m+1)n+m-1}{(m+2)n+m}\frac{\sqrt{2}S_{m+2}}{S_{m+3}}\frac{m+3}{m+1}k^{\frac{1}{n}}\right\},\end{split} (1.9) \displaystyle\begin{split}\alpha=&{V(\Omega)}/{(2\pi)^{n}},\,\,\,\rho=2(2\pi)^{-n}\sqrt{V(\Omega)I(\Omega)},\end{split} and a is defined by ( 2.15 ). In 2025, Ji and Xu [ 20 ] improved and generalized Ilyin’s eigenvalue estimates for any n\geq 2 under certain restrictive conditions. For any positive integer l\geq 1 , the Dirichlet eigenvalue problem of the poly-Laplacian is defined by the following equation (1.10) \begin{cases}(-\Delta)^{l}\bar{u}_{j}=\Lambda_{j}\bar{u}_{j}\ &\mathrm{in}\ \Omega,\\ \bar{u}_{j}=\frac{\partial\bar{u}_{j}}{\partial\nu}=\cdots=\frac{\partial^{l-1}\bar{u}_{j}}{\partial\nu^{l-1}}=0\ &\mathrm{on}\ \partial\Omega.\end{cases} If l=2 , equation ( 1.10 ) is called the clamped plate problem [ 7 , 8 , 18 , 19 , 20 ] , which governs motion of the vibration of a stiff plate. An important eigenvalue problem for the lower bound of the clamped plate problem is the famous Lord Rayleigh’s conjecture [ 21 , 22 ] . For the Brezin-Li-Yau type inequalities of the clamped plate problem, Cheng and Wei made important contributions [ 7 , 8 ] , and Ji and Xu [ 19 ] obtained an important result similar to ( 1.8 ). In [ 23 ] , Ku-Ku-Tan proposed the generalized Pólya’s Conjecture that the Dirichlet eigenvalues \Lambda_{k} of the poly-Laplacian satisfy inequalities \displaystyle\Lambda_{k}\geq \displaystyle\omega^{\frac{2l}{n}}_{n}\alpha^{-\frac{2l}{n}}k^{\frac{2l}{n}}. Recently, Ji and Xu [ 20 ] established a series of inequalities concerning this conjecture. Before we state our main results, we would like to discuss the following ways to obtain lower bounds for Dirichlet eigenvalues of the Laplacian and explain our methodology which helps us to obtain the improvements. Li-Yau’s way: Li-Yau [ 26 ] originally used the Fourier transform to covert the estimation of \sum_{i=1}^{k}\lambda_{i} to the upper bound M_{1} of certain decreasing radial rearrangement function F_{1}^{*}(|\xi|) . They proved the following key inequality \displaystyle\frac{n}{n+2}\left(\frac{n}{M_{1}\omega_{n-1}}\right)^{\frac{2}{n}}\left(\int_{\mathbb{R}^{n}}F^{*}_{1}(|\xi|)d\xi\right)^{\frac{n+2}{n}}\leq M_{2}, where M_{1} is the upper bound of F^{*}_{1} and M_{2} is the upper bound of \int_{\mathbb{R}^{n}}|\xi|^{2}F^{*}_{1}(|\xi|)d\xi . Using properties of the Fourier series and the rearrangement inequality, Li-Yau proved the upper bound of M_{1} is \alpha , then they obtained ( 1.3 ). Melas’s way: Melas [ 29 ] noticed that \displaystyle\bar{M}_{1}:=\int_{\mathbb{R}^{n}}|\xi|^{2}F^{*}_{1}(|\xi|)d\xi\leq\sum_{i=1}^{k}\lambda_{i}. Moreover, he proved that the upper bound for the derivative of F^{*}_{1}(|\xi|) is \rho . Then, Melas established a lower bound for a three-terms binary polynomial to obtain the lower bound for \bar{M}_{1} and proved ( 1.4 ). Later, Ji and Xu [ 19 ] found a better lower bound for this binary polynomial to control \bar{M}_{1} , then they improved ( 1.4 ) by giving ( 1.8 ). Consequently, if one can provide a sharper lower bound for \bar{M}_{1} or a sharper upper bound for M_{1} , the lower bound for \sum_{i=1}^{k}\lambda_{i} can be improved. These methods can be also used to investigate other eigenvalue problems [ 7 , 5 , 6 , 8 , 19 , 18 , 40 , 41 ] . In the Melas’s improvement of the Berezin-Li-Yau’s inequality and further improvements of Melas’s inequality (cf. [ 19 ] ), the key step is to give a lower bound for the following binary polynomial P(s,\tau):=ns^{n+2}-(n+2)\tau^{2}s^{n}+2\tau^{n+2}. Along this line, we get the best possible

Read the complete text