On the comparison principle for a nonlocal infinity Laplacian

Frida Fejne · 2025

Read On the comparison principle for a nonlocal infinity Laplacian by Frida Fejne (2025) 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-2501-04506.txt.

Introduction

Frida Fejne proves a uniqueness result for an equation built from a nonlocal version of the infinity Laplace operator, in a paper from 2025. Such operators measure extreme slopes rather than average ones. Fejne establishes a comparison principle for a bounded region with a continuous, non positive source term. That principle forces any two solutions to match, so the solution is unique.

Catalog record

CREATORFrida Fejne
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE43 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-04506

Browse the full library · About the Tower of Babel

Full text

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

On the comparison principle for a nonlocal infinity Laplacian Frida Fejne Abstract In this article, we prove the uniqueness of viscosity solutions to ℒ ∞ ​ u = f ​ in ​ Ω , \mathcal{L}_{\infty}u=f\textup{ in }\Omega, where ℒ ∞ \mathcal{L}_{\infty} denotes the nonlocal infinity Laplace operator, Ω \Omega a bounded domain, and f f a continuous functions such that f ≤ 0 f\leq 0 . Uniqueness is established through a comparison principle. 1 Introduction In this work, we consider uniqueness of viscosity solutions to Poisson’s equation for the nonlocal infinity Laplacian where the right-hand side is a function that does not change sign. More precisely, we let \Omega\subset\mathbb{R}^{n} be a bounded domain and consider the problem \mathcal{L}_{\infty}u=f\textup{ in }\Omega, (1.1) where f\in C(\Omega) is a bounded function such that f\leq 0 , and \mathcal{L}_{\infty}u denotes the nonlocal infinity Laplace operator which is defined in ( 1.5 ). Related problems have been studied extensively throughout the years. In 1967 in [ 2 ] Aronsson introduced the fundamental operator known as the infinity Laplacian: \Delta_{\infty}u(x):=D^{2}u(x)Du(x)\cdot Du(x)=\sum_{i,j=1}^{n}D^{2}_{ij}uD_{i}uD_{j}u. The equation \Delta_{\infty}u(x)=0 is the Euler-Lagrange equation when minimizing the supremum norm of the gradient of a function over a domain \Omega with some boundary data g . The solution to this problem is called an infinity harmonic function and when g is Lipschitz continuous, it will also be Lipschitz continuous. Furthermore, the infinity harmonic function with Lipschitz continuous boundary function g will have the least Lipschitz norm among functions with the same boundary data. Therefore this minimization problem is known as the minimal Lipschitz extension problem. In his classic paper [ 16 ] , from 1993, Jensen proved that the minimizing functions are equivalent to viscosity solutions of the homogeneous infinity Laplace equation and established uniqueness. Since then, many researcher have studied and contributed to the theory of infinity harmonic functions. For instance, in [ 3 ] Barles and Busca presented a new and different proof of Jensen’s theorem, which also worked for more general degenerate elliptic equations. For further references on the homogeneous infinity Laplace equation, see, e.g. [ 1 , 10 , 11 , 18 , 21 ] , but we note that it is very far from a complete list. Naturally, the infinity Laplace equation with inhomogeneous right-hand side has also been studied. In [ 23 ] , Lu and Wang worked with the Dirichlet problem \displaystyle\begin{cases}\Delta_{\infty}u=f&\textup{ in }\Omega,\\ u=g&\textup{ on }\partial\Omega.\end{cases} They established the existence and uniqueness of viscosity solutions under the assumption that the right-hand side f is continuous and strictly positive or strictly negative, that is, it stays away from zero. In addition, they also provided a counterexample of the uniqueness when f changes its sign. To our knowledge, the case where f\geq 0 , is still an open problem. The work of Lu and Wang was continued in [ 19 ] by Juutinen, Parviainen and Rossi. They worked with the gradient constraint problem \displaystyle\begin{cases}\textup{min}\{\Delta_{\infty}u(x),|Du(x)|-\chi_{D}\}=0&\textup{ in }\Omega,\\ u=g&\textup{ on }\partial\Omega.\end{cases} Here, \chi_{D} denotes the characteristic function of a set D\subset\Omega . Under the relatively mild regularity assumption that \overline{D}=\overline{\mathrm{int}\,D} they proved the uniqueness of viscosity solutions to the boundary value problem above. For more references on the inhomogeneous infinity Laplace equation, see, e.g. [ 4 , 9 , 15 , 25 ] . One nonlocal version of a fractional p -Laplace problem in a bounded domain \Omega can be defined as \displaystyle\begin{cases}(-\Delta_{p})^{s}u(x)=0&\textup{ in }\Omega,\\ u=g&\textup{ on }\partial\Omega,\end{cases} (1.2) where s\in(0,1) and (-\Delta_{p})^{s}u(x)=\lim_{\epsilon\to 0}\int_{\Omega\setminus B_{\epsilon(x)}}\frac{|u(x)-u(y)|^{p-2}(u(x)-u(y))}{|x-y|^{n+sp}}\ dy. (1.3) Here n denotes the dimension. We call this operator a fractional p -Laplace operator. We note that there is no standard definition when discussing this operator and the definitions in the literature varies, see e.g., [ 24 ] for a different version. As p tends to infinity, the limit problem in ( 1.2 ) becomes \displaystyle\begin{cases}(L_{\infty})^{\alpha}u(x)=0&\textup{ in }\Omega,\\ u=g&\textup{ on }\partial\Omega,\end{cases} (1.4) where we have used the notation \alpha instead of s , for \alpha\in(0,1] and (L_{\infty})^{\alpha}u(x):=\sup_{y\in\overline{\Omega}}\frac{u(y)-u(x)}{|x-y|^{\alpha}}+\inf_{y\in\overline{\Omega}}\frac{u(y)-u(x)}{|x-y|^{\alpha}}. Note that this equation is meaningful for \alpha=1 . This operator is called the Hölder infinity Laplacian. The reason behind this name is that when the boundary function g is Hölder continuous, more specifically g\in C^{0,\alpha}(\partial\Omega) , the solution u_{\infty} , to ( 1.4 ), has the smallest Hölder seminorm among functions with the same boundary data. Thus, the solution is the optimal Hölder extension. This problem and the corresponding non-homogeneous Dirichlet problem were first introduced in [ 8 ] where Chambolle et al. proved the existence of and some regularity results for solutions to (L_{\infty})^{\alpha}u=f . Note that this equation is, in general, not obtained as a limit of the p -Laplacian unless f=0 . In the case where f=0 they derived a representation formula for the solutions and proved the uniqueness through a comparison principle. In [ 20 ] Lindgren and Lindqvist worked on an eigenvalue problem involving the operator in ( 1.3 ) and functions with compact support. They also studied the passage to infinity, and after letting p\to\infty they obtained an expression involving (\mathcal{L}_{\infty})^{\alpha}u(x)=\sup_{y\in\mathbb{R}^{n}}\frac{u(y)-u(x)}{|x-y|^{\alpha}}+\inf_{y\in\mathbb{R}^{n}}\frac{u(y)-u(x)}{|x-y|^{\alpha}}, (1.5) for some \alpha\in(0,1] . This operator is called a nonlocal infinity Laplacian. It is also worth noting that the nonlocal infinity Laplacian can be obtained as a limit of more general fractional operators, see [ 13 ] for details. Furthermore, we note that the definition of the nonlocal infinity Laplacian operator may vary; see, e.g., [ 5 ] for a different version. In [ 14 ] the authors extended the problem in ( 1.2 ) to include functions with noncompact support and a more general right-hand side. They studied weak and viscosity solutions to \displaystyle\begin{cases}-(\mathcal{L}_{p})^{s}u(x)=f(x,u)&\textup{ in }\Omega,\\ u=g&\textup{ on }\mathbb{R}^{n}\setminus\Omega,\end{cases} where (\mathcal{L}_{p})^{s}u(x)=\lim_{\epsilon\to 0}\int_{\mathbb{R}^{n}\setminus B_{\epsilon}(x)}2\frac{|u(y)-u(x)|^{p-2}(u(y)-u(x))}{|x-y|^{n+sp}}\ dy, for different functions f on the right-hand side. Here 0<s<1 and \Omega is a Lipschitizian bounded domain. They also studied the limit equation after letting p tend to infinity and obtained an expression involving the nonlocal infinity Laplacian. For more related work on the fractional p -Laplacian and the nonlocal infinity Laplacian, see e.g., [ 6 , 7 , 12 ] . The operator in ( 1.5 ) is fundamental for our problem. We note that it depends on \alpha but since it will be clear from context what \alpha is, we will just denote it by \mathcal{L}_{\infty}u(x) . The nonlocal infinity Laplacian can be decomposed as \mathcal{L}_{\infty}u(x)=\mathcal{L}_{\infty}^{-}u(x)+\mathcal{L}_{\infty}^{+}u(x), where \displaystyle\mathcal{L}^{-}_{\infty}u(x):=\inf_{y\in\mathbb{R}^{n}}\frac{u(y)-u(x)}{|x-y|^{\alpha}}\quad\textup{and}\quad\mathcal{L}^{+}_{\infty}u(x):=\sup_{y\in\mathbb{R}^{n}}\frac{u(y)-u(x)}{|x-y|^{\alpha}}. Since the operator is not on divergence form there is no natural weak formulation of the problem in ( 1.1 ) and we therefore consider viscosity solutions, see Definition 2.1 . In this work, we show a comparison principle for viscosity solutions to ( 1.1 ) in the case where 0<\alpha\leq 1 . This is Theorem 3.1 . We are not considering existence, and therefore we do not work directly with the boundary function g . However, we impose some conditions on our functions outside the domain \Omega , namely that they are bounded on \mathbb{R}^{n}\setminus\Omega and that they have limits at infinity. Furthermore, we also assume that the right-hand side f\in C(\Omega) is bounded and does not switch sign, that is, f\leq 0 . The same results can also be obtained using similar methods if f is nonnegative. The proof of Theorem 3.1 uses ideas from [ 8 ] and

Read the complete text