Yang Eigenvalue Inequalities

• Yang eigenvalue inequalities are universal spectral gap inequalities that offer sharp, domain-independent bounds and recursion relations among eigenvalues for a variety of operators.
• They are derived using commutator, Bochner, and test-function methods, generalizing classical results such as the Payne–Pólya–Weinberger and Hile–Protter inequalities.
• The framework extends to weighted divergence forms, discrete graphs, higher-order, and subelliptic operators, providing significant insights in spectral geometry and eigenvalue estimation.

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 (Chen et al., 2013, Silva et al., 2023, Gomes et al., 2016, Huang et al., 2014, Huang et al., 2014, Filho, 2022).

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 (Silva et al., 2023, Gomes et al., 2016).

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 (Hua et al., 2017, Hua et al., 2020, Liu, 15 Jan 2026).

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 (Huang et al., 2014, Huang et al., 2014, Filho, 2022, Aribi et al., 2013).

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 $\gamma > 2$ (Karpukhin, 2019). 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 (Hua et al., 2017, Hua et al., 2020).

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 (Chen et al., 2013, Karpukhin, 2019).

---

References:

• "Payne-Polya-Weinberger, Hile-Protter and Yang's inequalities for Dirichlet Laplace eigenvalues on integer lattices" (Hua et al., 2017)
• "On the Yang-Yau inequality for the first Laplace eigenvalue" (Karpukhin, 2019)
• "Universal inequalities for Dirichlet eigenvalues on discrete groups" (Hua et al., 2020)
• "Inequalities for eigenvalues of the weighted Hodge Laplacian" (Chen et al., 2013)
• "Inequalities and bounds for the eigenvalues of the sub-Laplacian on a strictly pseudoconvex CR manifold" (Aribi et al., 2013)
• "Inequalities of Dirichlet eigenvalues for degenerate elliptic partial differential operators" (Huang et al., 2014)
• "Some New Inequalities of Dirichlet Eigenvalues for Laplace Operator with any Order" (Huang et al., 2014)
• "Inequalities for eigenvalues of fourth-order elliptic operators in divergence form on complete Riemannian manifolds" (Filho, 2022)
• "Some Eigenvalue Inequalities for the Schrödinger Operator on Integer Lattices" (Liu, 15 Jan 2026)
• "Eigenvalue estimates for a class of elliptic differential operators in divergence form" (Gomes et al., 2016)
• "Inequalities for eigenvalues of operators in divergence form on Riemannian manifolds isometrically immersed in Euclidean space" (Silva et al., 2023)
• "Eigenvalue inequalities for Klein-Gordon Operators" (Harrell et al., 2008)