New inequalities for eigenvalues of the Dirichlet Laplacian on the hyperbolic space

Yong Luo · 2026

Read New inequalities for eigenvalues of the Dirichlet Laplacian on the hyperbolic space by 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-2604-20343.txt.

Introduction

Luo proves new bounds for eigenvalues of the Dirichlet Laplacian on hyperbolic space, spurred by universal eigenvalue bounds. He checks Cheng's guess up to a small loss on two special bounded domain types in hyperbolic space. The work adds to the theory of universal bounds, showing the guessed bounds hold nearly exactly in these cases. Bounded domains of the two special kinds thus inherit eigenvalue bounds close to the guessed form.

Catalog record

CREATORYong Luo
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE20006
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.20343

Browse the full library · About the Tower of Babel

Full text

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

New inequalities for eigenvalues of the Dirichlet Laplacian on the hyperbolic space Yong Luo Abstract In this paper, motivated by study on universal inequalities for eigenvalues of the Dirichlet Laplacian, we prove some new inequalities for eigenvalues of the Dirichlet Laplacian on the hyperbolic space. In particular, we verify Cheng’s conjecture (Adv. Lect. Math. 37, 2017) up to loss of ϵ \epsilon for two special kinds of bounded domains in the hyperbolic space. AMS subject classifications: 58J50, 58E11, 35P15. Key words: Eigenvalues, universal inequality, the Dirichlet Laplacian, the hyperbolic space. 1 Introduction Let (M^{n},g) be an n -dimensional complete Riemannian manifold and \Omega a bounded connected domain in M . Let \Delta denote the Laplacian acting on functions on M . Eigenvalues of the Dirichlet Laplacian is stated as: \displaystyle\left\{\begin{array}[]{c}\Delta u=-\lambda u\text{ in }\Omega,\\ u|_{\partial\Omega}=0.\end{array}\right. Let \displaystyle 0<\lambda_{1}<\lambda_{2}\leq\lambda_{3}\leq\cdots denote the successive eigenvalues of ( 1 ). Here each eigenvalue is repeated according to its multiplicity. It is well-known that Weyl’s asymptotic formula holds (cf. [ 3 ] ): \displaystyle\lambda_{k}\sim\frac{4\pi^{2}}{(\omega_{n}Vol\Omega)^{\frac{2}{n}}}k^{\frac{2}{n}},\ k\to\infty, (1.4) where \omega_{n} is the volume of the unit ball in \mathbb{R}^{n} . Generally, if an inequality of eigenvalues requires no hypotheses on the geometric quantities of the domain \Omega (other than its dimension), it is called an universal inequality. Universal inequalities constitute an important research area in the study of the spectrum of \Delta and many works have been done during the past decades. Now we give a brief introduction of the main results. The study of universal inequalities for ( 1 ) was initiated by Payne, Pölya, and Weinberger [ 13 ] in 1956, they proved \displaystyle\lambda_{k+1}-\lambda_{k}\leq\frac{4}{kn}\sum_{i=1}^{k}\lambda_{i} for \Omega\subset\mathbb{R}^{n} (they originally proved the case n=2 , but their result can promoted to dimension n exactly by the same proof ). In 1980, Hile and Protter [ 11 ] proved \displaystyle\sum_{i=1}^{k}\frac{\lambda_{i}}{\lambda_{k+1}-\lambda_{i}}\geq\frac{kn}{4}. In 1991, Yang [ 15 ] (cf. [ 8 ] ) proved a sharp universal inequality: \displaystyle\sum_{i=1}^{k}\left(\lambda_{k+1}-\lambda_{i}\right)^{2}\leq\frac{4}{n}\sum_{i=1}^{k}\left(\lambda_{k+1}-\lambda_{i}\right)\lambda_{i}, (1.5) which has been called Yang’s first inequality by Ashbaugh (cf. [ 1 ] and [ 2 ] and so on). When \Omega is a bounded domain on \mathbb{S}^{n}(1) , Cheng and Yang [ 6 ] proved in 2005 that eigenvalues of ( 1 ) satisfy \displaystyle\sum_{i=1}^{k}\left(\lambda_{k+1}-\lambda_{i}\right)^{2}\leq\frac{4}{n}\sum_{i=1}^{k}\left(\lambda_{k+1}-\lambda_{i}\right)\left(\lambda_{i}+\frac{n^{2}}{4}\right), (1.6) by canonically embedded \mathbb{S}^{n}(1) into \mathbb{R}^{n+1} (see also [ 4 ] and [ 14 ] ). It is optimal since the above inequality becomes an equality for any k when \Omega=\mathbb{S}^{n}(1) . It is very natural to consider universal inequalities for eigenvalues of ( 1 ) when M^{n} is the hyperbolic space \mathbb{H}^{n}(-1) . If n=2 , by making use of estimates for eigenvalues of the Schrödinger like operator with a weight, Harrell and Michel [ 10 ] and Ashbaugh [ 2 ] have obtained several results. For general n , Cheng and Yang [ 9 ] proved in 2009 that eigenvalues of ( 1 ) satisfy \displaystyle\sum_{i=1}^{k}\left(\lambda_{k+1}-\lambda_{i}\right)^{2}\leqslant 4\sum_{i=1}^{k}\left(\lambda_{k+1}-\lambda_{i}\right)\left(\lambda_{i}-\frac{(n-1)^{2}}{4}\right). (1.7) It remains a challenging problem if one can improve the coefficient on the right hand side of (\ref{Cheng's 1}) from 4 to \frac{4}{n} (which is sharp by making use of a recursion formula of Cheng and Yang [ 8 ] and Weyl’s asymptotic formula ( 1.4 )). Actually we have a conjecture raised by Cheng [ 5 ] . Conjecture: Assume that \lambda_{i} are eigenvalues of ( 1 ) with (M^{n},g) being the hyperbolic space \mathbb{H}^{n}(-1) . Then have we \displaystyle\sum_{i=1}^{k}\left(\lambda_{k+1}-\lambda_{i}\right)^{2}\leqslant\frac{4}{n}\sum_{i=1}^{k}\left(\lambda_{k+1}-\lambda_{i}\right)\left(\lambda_{i}-\frac{(n-1)^{2}}{4}\right). Recently, motivated by Cheng’s conjecture, Luo and Zheng [ 12 ] studied eigenvalues of the Dirichlet Laplacian on the hyperbolic space \mathbb{H}^{n}(-1) , by viewing the hyperbolic space as a conformally flat Riemannian manifold. They proved Theorem 1.1 ( [ 12 ] ) . Assume that (M^{n},g) is the hyperbolic space \mathbb{H}^{n}(-1)=(\mathbb{R}_{+}^{n},\frac{dx_{1}^{2}+dx_{2}^{2}+\cdots+dx_{n}^{2}}{x_{n}^{2}}) , then eigenvalues \lambda_{i} of ( 1 ) satisfy \displaystyle\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})^{2}\leqslant\frac{\rho_{max}}{\rho_{min}}\frac{4}{n}\sum_{i=1}^{k}\left(\lambda_{k+1}-\lambda_{i}\right)\left(\lambda_{i}-\frac{n^{2}-2n-4}{4}\right), (1.8) where \rho_{max}:=\max_{\Omega}\frac{1}{x_{n}^{2}} and \rho_{min}:=\min_{\Omega}\frac{1}{x_{n}^{2}} . In this paper, by choosing new test functions, we obtain new inequalities for eigenvalues of the Dirichlet Laplacian on the hyperbolic space. Theorem 1.2 . Assume that (M^{n},g) is the hyperbolic space \mathbb{H}^{n}(-1)=(\mathbb{R}_{+}^{n},\frac{dx_{1}^{2}+dx_{2}^{2}+\cdots+dx_{n}^{2}}{x_{n}^{2}}) , then eigenvalues \lambda_{i} of ( 1 ) satisfy \displaystyle\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})^{2}\leqslant\frac{\rho_{max}}{\rho_{min}}\frac{4}{n}\sum_{i=1}^{k}\left(\lambda_{k+1}-\lambda_{i}\right)\left(\lambda_{i}-\frac{(n-1)^{2}}{4}\right). (1.9) Since (n-1)^{2}>n^{2}-2n-4 , we see that Theorem 1.2 improves Theorem 1.1 due to Luo and Zheng. As a direct corollary, we verify Cheng’s conjecture up to loss of \epsilon for a special kind of bounded domains. Corollary 1.3 . Assume that \Omega is a bounded domain in the hyperbolic space \mathbb{H}^{n}(-1)=(\mathbb{R}_{+}^{n},\frac{dx_{1}^{2}+dx_{2}^{2}+\cdots+dx_{n}^{2}}{x_{n}^{2}}) satisfying \frac{\rho_{max}}{\rho_{min}}\leq(1+\epsilon), then eigenvalues \lambda_{i} of ( 1 ) satisfy \displaystyle\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})^{2}\leqslant\frac{4(1+\epsilon)}{n}\sum_{i=1}^{k}\left(\lambda_{k+1}-\lambda_{i}\right)\left(\lambda_{i}-\frac{(n-1)^{2}}{4}\right). (1.10) By choosing different test functions, we verify Cheng’s conjecture up to loss of \epsilon for another special kind of bounded domains. Theorem 1.4 . Assume that (M^{n},g) is the hyperbolic space \mathbb{H}^{n}(-1)=(\mathbb{R}_{+}^{n},\frac{dx_{1}^{2}+dx_{2}^{2}+\cdots+dx_{n}^{2}}{x_{n}^{2}}) . Then for any 0<\epsilon\leq 1(n=2) or 0<\epsilon\leq 2(n\geq 3) , if \Omega is a bounded domain in \mathbb{H}^{n}(-1) satisfying \max_{x\in\Omega}\sum_{p=1}^{n-1}\frac{(\frac{x_{p}}{x_{n}})^{2}}{1+(\frac{x_{p}}{x_{n}})^{2}}\leq\frac{\epsilon^{2}}{1+\epsilon}, then eigenvalues \lambda_{i} of ( 1 ) satisfy \sum_{i=1}^{k}\left(\lambda_{k+1}-\lambda_{i}\right)^{2}\leq\frac{4(1+\epsilon)}{n}\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})(\lambda_{i}-\frac{(n-1)^{2}}{4}). Organization. In section 2 we collect some formulas from conformal geometry which will be useful in section 3 and in seciton 3 we give proofs of our main Theorem 1.2 and Theorem 1.4 . 2 Preliminaries Let (M,g_{0}) be an n -dimensional Riemannian manifold. Given a function h : M \rightarrow \mathbb{R} , we consider the conformal manifold (M,g) with the conformal metric g=e^{2h}g_{0} . Let \nabla^{0} , \nabla denote the Riemannian connections and \Delta^{0} , \Delta denote the Laplacian on (M,g_{0}) and (M,g) respectively. We denote \nabla^{0}F\cdot\nabla^{0}G=\langle\nabla^{0}F,\nabla^{0}G\rangle_{g_{0}},\ \nabla F\cdot\nabla G=\langle\nabla F,\nabla G\rangle_{g} for any functions F,G : \Omega \rightarrow \mathbb{R} . Then we have \displaystyle\nabla F\cdot\nabla G \displaystyle=\langle\nabla F,\nabla G\rangle_{g} \displaystyle=e^{2h}\left\langle e^{-2h}\nabla^{0}F,e^{-2h}\nabla^{0}G\right\rangle_{g_{0}} \displaystyle=e^{-2h}\langle\nabla^{0}F,\nabla^{0}G\rangle_{g_{0}} \displaystyle=e^{-2h}\nabla^{0}F\cdot\nabla^{0}G. (2.1) There is a known formula of the Laplacian under conformal transformation, which states that for a smooth function F : M \rightarrow \mathbb{R} , \displaystyle\quad\Delta F=e^{-2h}\left(\Delta^{0}F+(n-2)\nabla^{0}h\cdot\nabla^{0}F\right). (2.2) 3 Proofs of Theorem 1.2 and Theorem 1.4 In the following we will use a general inequality for eigenvalues of the Dirichlet Laplacian due to Cheng and Yang [ 7 ] . Proposition 3.1 ( [ 7 ] ) . Let \lambda_{i} be the i -th eigenvalue of the above eigenvalue problem 1 and u_{i} be the orthonormal eigenfunction corresponding to \lambda_{i} , i.e. \displaystyle\int_{M}u_{i}u_{i}=\delta_{ij},\int_{M}|\nabla u_{i}|^{2}=\lambda_{i}. Then, for any function f\in C^{3}(M)\cap C^{2}(\partial M) and any integer k , we have \displaystyle\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})^{2}\|u_{i}\nabla f\|^{2}\leq\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})\|2\nabla f\cdot\nabla u_{i}+u_{i}\Delta f\|^{2}, (3.1) where \|f\|^{2}=\int_{M}f^{2}. Let \mathbb{H}^{n}(-1) be the hyperbolic space with the Poincar \acute{e} upper half space model with Riemannian metric g , i.e. g=\frac{1}{x_{n}^{2}}(dx_{1}^{2}+\cdots+dx_{n}^{2}) . Proof of Theorem 1.2 . Let f_{p}=x_{p},p=1,\cdots,n-1 , using formulas ( 2 ) and ( 2.2 ) we have |\nabla f_{p}|^{2}=x_{n}^{2},\ \nabla f_{p}\cdot\nabla u_{i}=x_{n}^{2}\frac{\partial u_{i}}{\partial x_{p}},\ and\ \Delta f_{p}=0. Then we get \sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})^{2}\int_{\Omega}u_{i}^{2}x_{n}^{2}\leq 4\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})\int_{\Omega}x_{n}^{4}(\frac{\partial u_{i}}{\partial x_{p}})^{2}. Therefore we obtain \min_{\Omega}x_{n}^{2}\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})^{2}\int_{\Omega}u_{i}^{2}\leq\max_{\Omega}x_{n}^{2}4\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})\int_{\Omega}x_{n}^{2}(\frac{\partial u_{i}}{\partial x_{p}})^{2}, which implies \displaystyle\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})^{2}\leq\frac{\max_{\Omega}x_{n}^{2}}{\min_{\Omega}x_{n}^{2}}4\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})\int_{\Omega}x_{n}^{2}(\frac{\partial u_{i}}{\partial x_{p}})^{2}, (3.2) for each p=1,\cdots,n-1. Let f_{n}=\ln x_{n}, using formulas ( 2 ) and ( 2.2 ) we have |\nabla f_{n}|^{2}=1,\ \nabla f_{n}\cdot\nabla u_{i}=x_{n}\frac{\partial u_{i}}{\partial x_{n}},\ and\ \Delta f_{n}=1-n. Then we get \sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})^{2}\int_{\Omega}u_{i}^{2}\leq\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})\int_{\Omega}\left(2x_{n}\frac{\partial u_{i}}{\partial x_{n}}+u_{i}(1-n)\right)^{2}. Note that \displaystyle\int_{\Omega}\left(2x_{n}\frac{\partial u_{i}}{\partial x_{n}}+u_{i}(1-n)\right)^{2} \displaystyle= \displaystyle\int_{\Omega}\left(4x_{n}^{2}(\frac{\partial u_{i}}{\partial x_{n}})^{2}+4(1-n)\nabla x_{n}\cdot\nabla u_{i}u_{i}+(n-1)^{2}u_{i}^{2}\right) \displaystyle= \displaystyle\int_{\Omega}\left(4x_{n}^{2}(\frac{\partial u_{i}}{\partial x_{n}})^{2}+2(1-n)\nabla x_{n}\cdot\nabla u_{i}^{2}+(n-1)^{2}u_{i}^{2}\right) \displaystyle= \displaystyle\int_{\Omega}\left(4x_{n}^{2}(\frac{\partial u_{i}}{\partial x_{n}})^{2}-2(1-n)\Delta x_{n}u_{i}^{2}+(n-1)^{2}u_{i}^{2}\right) \displaystyle= \displaystyle\int_{\Omega}\left(4x_{n}^{2}(\frac{\partial u_{i}}{\partial x_{n}})^{2}-(n-1)^{2}u_{i}^{2}\right), we obtain \displaystyle\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})^{2} (3.3) \displaystyle\leq \displaystyle 4\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})\left(\int_{\Omega}x_{n}^{2}(\frac{\partial u_{i}}{\partial x_{n}})^{2}-\frac{(n-1)^{2}}{4}\right) \displaystyle\leq \displaystyle\frac{\max_{\Omega}x_{n}^{2}}{\min_{\Omega}x_{n}^{2}}4\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})\left(\int_{\Omega}x_{n}^{2}(\frac{\partial u_{i}}{\partial x_{n}})^{2}-\frac{(n-1)^{2}}{4}\right). Summing p from 1 to n-1 in inequality ( 3.2 ) and inequality ( 3.3 ) we get \displaystyle n\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})^{2} \displaystyle\leq \displaystyle\frac{\max_{\Omega}x_{n}^{2}}{\min_{\Omega}x_{n}^{2}}4\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})\left(\int_{\Omega}|\nabla u_{i}|^{2}-\frac{(n-1)^{2}}{4}\right) \displaystyle= \displaystyle\frac{\max_{\Omega}x_{n}^{2}}{\min_{\Omega}x_{n}^{2}}4\sum_{i=1}^{k}(\lambda_{k+1}-\lambda_{i})\left(\lambda_{i}-\frac{(n-1)^{2}}{4}\right), where in the above inequality we

Read the complete text