🔢 mathematics

Toward Pólya's Conjecture: Improving the Individual Li-Yau Bound via Energy Orthogonality

This paper establishes two complementary lower-bound mechanisms for individual Dirichlet Laplacian eigenvalues on open sets in $\mathbb{R}^n$ ( $n\geq2$ ) by utilizing energy orthogonality and spectral deficit retention to strictly improve upon the individual consequences of the Li-Yau sum inequality while preserving Weyl scaling.

Original authors: Yifan Wang, Hehu Xie

Published 2026-09-16

📖 5 min read🧠 Deep dive

CC0 1.0

Original authors: Yifan Wang, Hehu Xie

Original paper dedicated to the public domain under CC0 1.0 (http://creativecommons.org/publicdomain/zero/1.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the quiet mathematics of shapes, there is a fundamental question about how energy behaves when trapped inside a container. Imagine a drumhead stretched tight over a circular frame; when struck, it vibrates at specific, distinct pitches. These pitches are not random; they are determined entirely by the size and shape of the drum. In the language of physics and mathematics, these pitches are called eigenvalues, and the container is a domain. For over a century, mathematicians have sought a simple rule to predict the lowest possible pitch for any given shape, regardless of its complexity. A famous conjecture suggests that the pitch is always at least as high as it would be for a perfect circle of the same area. While this idea has been proven for some special shapes, it remains unproven for the general case. The challenge lies in finding a lower limit that is both accurate and universal, one that holds true for every single vibration mode, from the deepest hum to the highest squeal, without needing to know the intricate details of the boundary.

Two researchers, Yifan Wang and Hehu Xie, have taken a significant step toward solving this puzzle by refining the tools used to estimate these pitches. Their work focuses on improving a known mathematical limit, specifically for the individual notes of a vibrating system, rather than just the average of many notes. They started with a well-established method that provides a safe, conservative estimate for these pitches. This older method works like a broad net, catching the general behavior of the vibrations but leaving some room for error because it treats all frequencies as if they were equally likely to occur. The researchers realized that this approach missed a crucial piece of information: the way the vibrations interact with each other. By looking at how the energy of these vibrations is distributed across different frequencies, they discovered a hidden constraint. Just as a crowded room has a limit to how many people can stand in a small space, the vibrations have a limit to how densely they can pack into a specific frequency range.

The team developed a new way to calculate this limit by combining two different types of information. The first is a standard measure of how much energy is present, which acts like a ceiling on how many vibrations can exist. The second, and more novel, part of their discovery comes from the fact that these vibrations are orthogonal, meaning they are independent of one another in a specific mathematical sense. This independence creates a second, stricter limit that depends on the frequency itself. At higher frequencies, this new limit becomes much tighter, effectively squeezing the possible values of the vibrations more than the old method allowed. By using a principle similar to filling a bathtub with water, where the water level rises to fill the available space up to a certain point, they calculated the minimum possible pitch for any given number of vibrations. This calculation works for any shape with a finite area, requiring no special assumptions about the smoothness of the edges.

Their findings show that the old estimate was indeed too low. The new method provides a strictly higher lower bound, meaning the true pitch is guaranteed to be higher than previously thought. In two-dimensional space, this improvement amounts to a roughly 7.66 percent increase in the estimated value, a significant gain in the world of precise mathematical constants. The researchers also created a more advanced version of their method for shapes with jagged or irregular boundaries, known as Lipschitz domains. This hybrid approach incorporates a detailed count of how many vibrations exist below a certain energy level, using a specific counting technique developed by other mathematicians. This allows the new bound to adapt to the geometry of the container, offering an even tighter estimate when the shape is known. However, the authors are careful to note that while their method improves upon the standard universal limit, it does not yet reach the ultimate goal of the famous conjecture, which would predict the pitch with perfect accuracy for every shape.

The paper also clarifies what their work does not do. It does not claim to have solved the entire conjecture for all shapes, nor does it provide a better estimate for the sum of all vibrations combined, where the old method remains the sharpest known tool. The improvement is specific to the individual notes. Furthermore, the researchers demonstrate that their new bound is not automatically stronger than every other specialized estimate that exists for specific, simple shapes like perfect circles or squares. For those special cases, other methods already provide the exact answer. Instead, their contribution is a robust, universal improvement that works everywhere, filling a gap in our understanding of how energy is distributed in vibrating systems. By proving that the vibrations must obey a stricter frequency-dependent rule, they have tightened the mathematical net, bringing us closer to the precise truth of how nature vibrates, one note at a time.

Technical Summary: Toward Pólya's Conjecture: Improving the Individual Li–Yau Bound via Energy Orthogonality

Problem Statement
The paper addresses the problem of establishing lower bounds for individual Dirichlet eigenvalues, $\lambda_k$ , of the Laplacian on an open set $\Omega \subset \mathbb{R}^n$ ( $n \ge 2$ ) with finite positive measure $V$ . While Weyl's law provides the asymptotic behavior $\lambda_k \sim C_n V^{-2/n} k^{2/n}$ , Pólya's conjecture posits that this term serves as a strict lower bound for every $k$ . The classical Li–Yau inequality provides a sharp bound for the sum of the first $k$ eigenvalues, $\sum_{j=1}^k \lambda_j \ge \frac{n}{n+2} C_n V^{-2/n} k^{1+2/n}$ . However, the immediate individual consequence derived from this sum, $\lambda_k \ge \frac{n}{n+2} C_n V^{-2/n} k^{2/n}$ , is known to be suboptimal compared to the conjectured coefficient of 1. The authors aim to improve this specific individual coefficient without relying on additional geometric assumptions (such as convexity) or boundary regularity, while acknowledging that the sharp leading coefficient for the sum inequality remains unchanged.

Methodology
The authors develop two complementary mechanisms to refine the lower bound:

Energy Orthogonality and Fourier Envelope:
The standard Li–Yau argument utilizes the $L^2$ Bessel inequality to establish a constant upper bound on the Fourier density of the spectral projection, $F_k(\xi) = \sum_{j=1}^k |\hat{\phi}_j(\xi)|^2$ . The authors introduce a second constraint derived from the orthogonality of the gradients of eigenfunctions (energy orthogonality). By expressing a plane wave as the divergence of a specific vector field and utilizing the weak formulation of the eigenvalue equation, they derive a frequency-dependent decay cap: $F_k(\xi) \le \frac{V}{(2\pi)^n} \frac{\lambda_k}{|\xi|^2}$ .
Combining this with the constant cap yields a radial envelope $h(|\xi|) = \frac{V}{(2\pi)^n} \min\{1, \frac{\lambda_k}{|\xi|^2}\}$ . The authors then apply a "bathtub principle" (a rearrangement inequality) to find the density $f(\xi)$ below this envelope that minimizes the second moment $\int |\xi|^2 f(\xi) d\xi$ subject to the mass constraint $\int f(\xi) d\xi = k$ . This minimization problem is solved explicitly for dimensions $n \ge 3$ and $n=2$ .
Hybrid Spectral Closure via Riesz Means:
The second mechanism addresses the "spectral deficit" discarded when bounding the sum $\sum_{j=1}^k \lambda_j$ by $k\lambda_k$ . The authors utilize the exact identity $S_k = k\lambda_k - R_1(\lambda_k)$ , where $R_1$ is the first Riesz mean (the integral of the counting function). By inserting a lower bound for the counting function $N_\Omega(t)$ into this identity, they construct a scalar equation $\Phi_k(L) = M_k(L) + J_k(L) - kL = 0$ . Here, $M_k(L)$ is the minimal moment derived from the envelope, and $J_k(L)$ is the integral of a truncated counting lower bound. The unique positive root of this strictly decreasing function provides a refined lower bound for $\lambda_k$ . This framework is implemented explicitly on bounded Lipschitz domains using the quantitative counting estimates of Jiang–Lin.

Key Contributions and Results

Universal Coefficient Improvement:
The primary result is a strictly improved individual lower bound valid for any open set of finite positive measure, without boundary regularity assumptions. The bound takes the form:
$\lambda_k \ge c_n C_n V^{-2/n} k^{2/n}$
where the coefficient $c_n$ satisfies $\frac{n}{n+2} < c_n < 1$ .
- For $n \ge 3$ , $c_n$ is determined by the unique root $x_n > 1$ of $(n-2)x^n - n x^{n-2} + \frac{8}{n+2} = 0$ , with $c_n = K_n^{-2/n}$ .
- For $n=2$ , $c_2 = K_2^{-1}$ where $K_2 = 1 + \log y_2$ and $y_2$ is the unique solution to $y - \log y = 3/2$ .
- Numerical values show a strict improvement over the Li–Yau coefficient: $c_2 \approx 0.5383$ (a $\approx 7.66\%$ gain over $0.5 $),$ c_3 \approx 0.6222 $(a$ \approx 3.71%$ gain over $0.6$), with the relative gain decreasing as dimension increases.
Hybrid Refinement on Lipschitz Domains:
On bounded Lipschitz domains, the paper constructs a "hybrid" bound $L_k^{hyb}$ by integrating the Jiang–Lin counting estimates into the Riesz-mean identity.
- Theorem 5.3 establishes that $L_k^{hyb}$ is the unique root of a monotone scalar equation and satisfies $L_k^{(0)} \le L_k^{hyb} < W_k$ (where $W_k$ is the Pólya term).
- The paper provides necessary and sufficient criteria (Equation 5.13) to determine when this hybrid bound strictly exceeds both the volume-only baseline and any other independent lower bound.
- The authors demonstrate that while this hybrid bound can offer significant gains (e.g., $\approx 15.3\%$ over the maximum of separate bounds in a numerical example), it does not automatically dominate all geometric estimates (such as those for specific domains like balls) or the full strength of sum inequalities.

Significance and Claims
The paper claims that its results represent a strict improvement over the individual consequence of the Li–Yau inequality for every index $k$ and every open set, preserving the correct Weyl scaling. The improvement arises from retaining information about energy orthogonality that is lost in the standard constant-cap Fourier argument.

The authors are explicit about the limitations of their findings:

No Improvement to Sum Inequality: The improved individual coefficient $c_n$ is strictly less than 1; thus, summing these individual bounds does not recover the sharp leading coefficient of the Li–Yau sum inequality.
No Universal Dominance: The hybrid bound does not universally dominate stronger estimates available under specific geometric assumptions (e.g., Pólya's inequality for balls, isoperimetric bounds, or high-frequency asymptotics with epsilon-loss).
Modest Scope: The paper positions itself as providing a transparent "baseline" and a compatible framework for incorporating additional geometric or spectral data (via the hybrid mechanism), rather than replacing existing specialized estimates. The authors note that reaching the conjectured Pólya coefficient of 1 would require stronger moment constraints or different spectral closures beyond the current framework.

In summary, the work advances the understanding of individual eigenvalue bounds by refining the Fourier envelope through energy orthogonality and offering a rigorous, monotone framework for integrating counting estimates, while carefully delineating the boundaries of these improvements relative to the broader landscape of spectral geometry.

More like this