---
title: Yang Eigenvalue Inequalities
url: https://www.emergentmind.com/topics/yang-eigenvalue-inequalities
type: topic
---

# Yang Eigenvalue Inequalities

Yang eigenvalue inequalities are "universal" spectral gap inequalities for Dirichlet-type eigenvalue problems, first established for the Laplacian on bounded domains and subsequently generalized to a broad range of elliptic, subelliptic, and discrete operators. These inequalities — and their extensions in various geometric, analytic, and combinatorial settings — delineate sharp domain-independent upper bounds on eigenvalue gaps and provide recursion relations among the sequence of eigenvalues. Their ubiquity stems from algebraic identities derived via commutator, Bochner, or test-function methods, and they offer a unified framework including the classical Payne–Pólya–Weinberger and Hile–Protter inequalities as corollaries.

## 1. Classical Formulation and Main Inequalities

The archetypal setting is the Dirichlet Laplacian on a bounded domain $\Omega\subset\mathbb{R}^n$, where eigenvalues are ordered $0<\lambda_1\le\lambda_2\le\lambda_3\le\dots$. H. C. Yang's first and second inequalities take the forms:
\[
\sum_{i=1}^{k} (\lambda_{k+1} - \lambda_i)^2 \le \frac{4}{n} \sum_{i=1}^{k} (\lambda_{k+1} - \lambda_i)\lambda_i
\]
\[
\lambda_{k+1} \leq \left(1 + \frac{4}{n}\right) \frac{1}{k} \sum_{i=1}^k \lambda_i
\]
These bounds are sharp in that the constants cannot be improved without restricting the class of domains.

Derivations rely on trace identities and test-function constructions exploiting the Euclidean structure, particularly using the coordinate functions $x_j$ as multipliers, leading to specific commutator estimates. Analogues for higher-order operators, elliptic operators with variable coefficients, and various geometric generalizations maintain the quadratic-in-gap structure of the inequalities, with explicit dependence on dimension and, where appropriate, geometric quantities such as mean curvature or tensor ellipticity bounds [1312.0218, 2305.12024, 1607.00066, 1405.0688, 1405.0690, 2201.06453].

## 2. Generalizations to Divergence-Form and Weighted Operators

Subsequent work generalized Yang’s inequalities to second-order elliptic operators in weighted divergence form on Riemannian manifolds. Consider the Dirichlet problem for $L_{\eta,T} u = \operatorname{div}_\eta(T(\nabla u)) = \operatorname{div}(T(\nabla u)) - (\nabla\eta, T(\nabla u))$, with $T$ a symmetric positive-definite $(1,1)$-tensor and $\eta$ a drift function. Eigenvalues $\lambda_1\le\lambda_2\le\dots$ satisfy:

\[
\sum_{i=1}^k (\lambda_{k+1}-\lambda_i)^2 \le \frac{1}{C(\varepsilon, n)} \sum_{i=1}^k (\lambda_{k+1}-\lambda_i)\bigl(\lambda_i + \Psi(\mathrm{geometry,\, T,\, \eta})\bigr)
\]
where $C(\varepsilon, n)$ is determined by the lower ellipticity bound $\varepsilon$ of $T$ and the manifold's dimension $n$, and $\Psi$ encodes extrinsic and analytic data: generalized mean curvature, drift, and divergence terms [2305.12024, 1607.00066].

The constants and geometric correction terms are explicit. For $T = I$, $\eta \equiv 0$ (classical Laplacian), the inequalities reduce to Yang’s original forms.

## 3. Discrete and Graph-Theoretic Settings

On discrete domains, such as subsets of the integer lattice $\mathbb{Z}^n$, a normalized combinatorial Laplacian $\Delta_\Omega$ is used. For finite $\Omega\subset\mathbb{Z}^n$, eigenvalues $0 < \lambda_1\le\dots\le \lambda_N\le 2$ satisfy the discrete Yang inequality:

\[
\sum_{i=1}^k (\lambda_{k+1} - \lambda_i)^2 (1-\lambda_i) \le \frac{4}{n} \sum_{i=1}^{k} (\lambda_{k+1} - \lambda_i)\lambda_i
\]
with further refinements and companion inequalities of Hile–Protter, PPW type, and extensions to general Schrödinger operators with variable weights:

\[
\sum_{i=1}^{k} (\lambda_{k+1}-\lambda_i)\left[ \lambda_{k+1} \alpha_i - \lambda_i (\alpha_i + \tfrac{4}{n}) \right] \le 0,
\]
where $\alpha_i$ encode the effect of weighting and the bipartite graph structure [1710.05799, 2007.13157, 2601.10523].

The combinatorial structure introduces spectrum symmetry and modified gap weights, with bipartiteness leading to additional $(1-\lambda_i)$ terms not present in the continuum.

## 4. Extensions to Higher-Order, Degenerate, and Subelliptic Operators

Universal Yang-type inequalities hold for higher-order Laplacians $(−\Delta)^\ell$ and degenerate elliptic or subelliptic operators (e.g., sub-Laplacians on stratified groups, Greiner-type operators, CR-manifolds). For the $\ell$-th power of the Laplacian:

\[
\sum_{i=1}^k (\lambda_{k+1} - \lambda_i)^2 \leq \frac{4\ell(n + 2\ell - 2)}{n^2} \sum_{i=1}^k (\lambda_{k+1} - \lambda_i)\lambda_i,
\]
and for sum-of-squares operators on Carnot or CR structures, the dimension $n$ is replaced by the rank of the horizontal bundle, and appropriate geometric corrections are included [1405.0690, 1405.0688, 2201.06453, 1301.6493].

The commutator technique generalizes: the key input is a commutator identity involving vector fields satisfying Hörmander’s condition and their associated domains.

## 5. Yang–Yau Inequality and Spectral Geometry on Closed Surfaces

P. C. Yang and S.-T. Yau established a bound for the first Laplace eigenvalue on a closed orientable Riemannian surface of genus $\gamma$ and total area $A$:

\[
\lambda_1\, A \leq 8\pi \left\lfloor \frac{\gamma+3}{2} \right\rfloor,
\]
with equality for $\gamma = 0$ (sphere) and strict inequality for $\gamma > 2$ [1902.03473]. This result is sharp in prescribed cases and has deep links to algebraic geometry via special divisors and harmonic maps into spheres. Its strictness for $\gamma > 2$ is established via Brill–Noether theory and minimal surface theory.

## 6. Methodologies: Commutator, Integration by Parts, and Recursion

Proofs exploit clever constructions of trial functions (typically coordinate-based or coming from isometric immersions/eigenmaps), used in combination with integration by parts or commutator identities, for instance:
- The commutator method for the Laplacian: $[A, x_j]u = -2\, \partial_{x_j}u$.
- Bochner-type identities and trace inequalities: $\sum_{i=1}^k (\lambda_{k+1}-\lambda_i)^2 \le \cdots$ via summing Rayleigh quotients.
- Algebraic Chebyshev-type rearrangement inequalities to handle more general sums and weights.

Recursive (Cheng–Yang type) formulations yield secondary inequalities, e.g., bounds for $\lambda_{k+1}$ in terms of averages of $\lambda_1, ..., \lambda_k$.

## 7. Optimality, Equality Cases, and Open Problems

The constants appearing in Yang-type inequalities are generally believed to be optimal, coinciding with scaling limits given by the Weyl law. Equality (or near equality) occurs only in highly constrained settings—typically the $1$-dimensional Dirichlet interval or small, highly symmetric domains (e.g., spheres for $\lambda_1 A = 8\pi$). In the discrete setting, the presence of bipartite/symmetry factors, or the sharp constants for trees and lattices, remain largely optimal but generic domains do not achieve equality [1710.05799, 2007.13157].

Open problems include finding geometric or combinatorial refinements to remove residual weights (like $(1-\lambda_i)$), the study of non-Laplacian operators (e.g., fractional Laplacians, weighted Hodge Laplacians), and classification of extremal domains or metrics, especially on higher genus surfaces and other topologically non-trivial manifolds [1312.0218, 1902.03473].

---

**References:**

- "Payne-Polya-Weinberger, Hile-Protter and Yang's inequalities for Dirichlet Laplace eigenvalues on integer lattices" [1710.05799]
- "On the Yang-Yau inequality for the first Laplace eigenvalue" [1902.03473]
- "Universal inequalities for Dirichlet eigenvalues on discrete groups" [2007.13157]
- "Inequalities for eigenvalues of the weighted Hodge Laplacian" [1312.0218]
- "Inequalities and bounds for the eigenvalues of the sub-Laplacian on a strictly pseudoconvex CR manifold" [1301.6493]
- "Inequalities of Dirichlet eigenvalues for degenerate elliptic partial differential operators" [1405.0688]
- "Some New Inequalities of Dirichlet Eigenvalues for Laplace Operator with any Order" [1405.0690]
- "Inequalities for eigenvalues of fourth-order elliptic operators in divergence form on complete Riemannian manifolds" [2201.06453]
- "Some Eigenvalue Inequalities for the Schrödinger Operator on Integer Lattices" [2601.10523]
- "Eigenvalue estimates for a class of elliptic differential operators in divergence form" [1607.00066]
- "Inequalities for eigenvalues of operators in divergence form on Riemannian manifolds isometrically immersed in Euclidean space" [2305.12024]
- "Eigenvalue inequalities for Klein-Gordon Operators" [0810.0059]

Source: https://www.emergentmind.com/topics/yang-eigenvalue-inequalities