A note on Liouville type theorem of elliptic inequality $Δu+u^σ\leq 0$ on Riemannian manifolds

Hui-Chun Zhang

TL;DR

This paper establishes a new, more general condition for a Liouville-type theorem for a specific type of elliptic inequality on Riemannian manifolds. Instead of relying on a pointwise bound on the volume growth of geodesic balls, the authors show that an integral condition on this volume growth is sufficient to guarantee that any non-negative solution to the inequality must be identically zero. This work extends previous results and provides a deeper understanding of the relationship between geometric properties of manifolds and the behavior of solutions to elliptic equations.

What Problem Does This Paper Address?

The paper tackles a fundamental question in the study of partial differential equations on manifolds: under what geometric conditions can we guarantee that certain solutions must be trivial (i.e., zero)? Specifically, it focuses on the elliptic inequality $\Delta u(x) + u^\sigma(x) \leq 0$, where $\Delta$ is the Laplace-Beltrami operator, $u$ is a non-negative function, $\sigma > 1$, and the inequality holds on a complete, non-compact Riemannian manifold $M$.

A "Liouville-type theorem" in this context means that any non-negative weak solution to this inequality must be identically zero. Previous work by Grigor'yan and Sun (and others) showed that a pointwise upper bound on the volume growth of geodesic balls (specifically, $V(r) < Cr^\frac{2\sigma}{\sigma-1}$ for large $r$) was sufficient for such a theorem. However, the authors of this paper were inspired by a related result for parabolicity (where the inequality is $\Delta u \leq 0$) which uses an integral condition on volume growth. They sought to generalize the Liouville theorem for the inequality $\Delta u + u^\sigma \leq 0$ by replacing the pointwise condition with a more flexible integral condition. This is important because integral conditions can sometimes be more natural to work with or might capture a broader class of manifolds.

Methods

The core of the paper's method involves constructing specific "test functions" and using them to derive integral estimates for the solution $u$. The approach is largely based on the techniques developed in the Grigor'yan-Sun paper [10], which provides useful a priori estimates.

  1. A Priori Estimates: The paper starts by recalling a crucial lemma (Lemma 2.1) from Grigor'yan-Sun. This lemma provides inequalities relating integrals of $|\nabla u|^2$ and $u^2$ to integrals involving a test function $\phi$ and its derivatives, weighted by a parameter $t$. These estimates depend on the parameter $\sigma$ and a related parameter $s = 8\sigma/(\sigma-1)$.
  2. Construction of Test Functions: The authors then construct a family of specific test functions, denoted by $\phi_{t,n}(x)$. These functions are designed to "probe" the behavior of the solution $u$ in different regions of the manifold, particularly as the radius of geodesic balls grows. They are modifications of functions used in previous work and are Lipschitz continuous with compact support.
  3. Integral Estimates and Limits: Using these test functions, the paper derives integral estimates for $u$ over various regions of the manifold. A key step involves calculating limits of these estimates as $n \to \infty$ (which effectively makes the test function "larger" in support) and as a parameter $t \to 0^+$. These limits are carefully analyzed using the volume growth function $V(r)$.
  4. Connecting to Volume Growth: The crucial step is to relate these integral estimates to the volume growth condition. The paper shows that if the integral condition on $V(r)$ (Theorem 1.2, condition (1.3)) is satisfied, then the derived integral estimates for $u$ must force $u$ to be zero. This is achieved by choosing a sequence of parameters $t_a$ that satisfy the integral condition and then taking limits.
  5. Comparison with Pointwise Condition: The paper also explicitly shows that the pointwise condition from Grigor'yan-Sun (Theorem 1.1, condition (1.2)) implies the new integral condition (Theorem 1.2, condition (1.3)), demonstrating that their result is indeed a generalization.

Datasets

Not discussed in the paper. This is a theoretical paper focused on differential geometry and analysis, not empirical data.

Key Insights and Ideas

  • Integral vs. Pointwise Conditions: The most significant insight is demonstrating that an integral condition on volume growth is sufficient for a Liouville-type theorem, generalizing the previous pointwise condition. This suggests that the "average" behavior of volume growth can be as powerful as its "worst-case" pointwise behavior for certain analytical problems.
  • Flexibility of Integral Conditions: Integral conditions are often more robust and can encompass a wider range of geometric scenarios than strict pointwise bounds. This work opens the door to applying Liouville theorems in situations where pointwise bounds might fail.
  • Connection to Parabolicity: The inspiration from Varopoulos-Grigor'yan's criterion for parabolicity highlights a deep connection between different types of elliptic and parabolic inequalities and their relationship to the geometric properties of the underlying manifold.
  • Power of Test Functions: The paper showcases the effectiveness of carefully constructed test functions in deriving sharp analytical estimates on manifolds, a common and powerful technique in geometric analysis.

Key Results

The main result is Theorem 1.2: If $M$ is a complete, non-compact Riemannian manifold and its volume growth function $V(r)$ satisfies the integral condition $$ \lim_{t \to 0^+} \int_{t}^{\infty} \frac{V(r)}{r^{\sigma-1+t}} dr < \infty $$ for $\sigma > 1$, then any non-negative weak solution to $\Delta u + u^\sigma \leq 0$ on $M$ must be identically zero.

The paper also proves that the pointwise condition $V(r) < Cr^{\frac{2\sigma}{\sigma-1}}$ (from Theorem 1.1) implies this new integral condition, thus establishing the generalization.

Limitations and Future Work

  • Completeness and Non-compactness: The theorems are stated for complete, non-compact Riemannian manifolds. It would be interesting to explore what happens on compact manifolds or manifolds with boundary, or manifolds that are not necessarily complete.
  • Specific Inequality: The results are specific to the inequality $\Delta u + u^\sigma \leq 0$. Extending these integral conditions to other types of elliptic inequalities or systems would be a natural next step.
  • Sharpness of the Condition: While the integral condition is shown to be sufficient, its necessity is not addressed. Determining if this integral condition is also necessary for the Liouville-type theorem would be valuable.
  • Curvature: The theorems are curvature-free, which is a significant achievement. However, exploring how curvature might influence the constants or the precise form of the integral condition could offer further insights.

Summary updated

Abstract

Let $σ&gt;1$ and let $M$ be a complete Riemannian manifold. In a recent work [9], Grigor$^{\prime}$yan and Sun proved that a pointwise upper bound of volume growth is sufficient for uniqueness of nonnegative solutions of elliptic inequality $$(*)\quad\qquad\qquad\qquad Δu(x)+u^σ(x)\leq 0,\qquad x\in M.\quad\qquad \qquad\qquad $$ In this note, we improve their result to that an \emph{integral condition} on volume growth implies the same uniqueness of ($*$). It is inspired by the well-known Varopoulos-Grigor$^{\prime}$yan's criterion for parabolicity of $M$.
BibTeX
@misc{zhang2014note,
  author = {Hui-Chun Zhang},
  title = {{A note on Liouville type theorem of elliptic inequality $Δu+u^σ\textbackslash\{\}leq 0$ on Riemannian manifolds}},
  year = {2014},
  eprint = {1404.5805},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1404.5805},
}

Loading related content...

Figure