1
$\begingroup$

Suppose $u$ is a sign changing classical solution of the fractional Laplacian $$ (-\Delta) ^{s} u = 0 \; \text{in } \Omega; u=g \text{ in } \mathbb R^N -\Omega .$$ (a)Is it true that $\|u\|_{L^{\infty}(\mathbb R^N)}\leq \|g\|_{L^{\infty}(\mathbb R^N -\Omega)}$ when $s\in (0, 1).$

(b) Is this also true for the operator $(-\Delta) ^{s}+I$ where $I$ is the identity.

$\endgroup$
5
  • 1
    $\begingroup$ Given some minimal regularity of $u$ — of course yes! After all, $u$ is given as an integral of $g$ against a (sub-)probability measure, the harmonic measure for $(-\Delta)^s$ (also called the $\alpha$-harmonic measure or the Poisson kernel for $(-\Delta)^s$). $\endgroup$ Aug 22, 2019 at 20:56
  • $\begingroup$ If we suppose $u$ to be Holder continuous on $\mathbb R^N.$ I am aware of the notion of Poisson kernel. $\endgroup$
    – GabS
    Aug 22, 2019 at 21:31
  • $\begingroup$ If your domain is nice -- Lipschitz for example -- it is sufficient to assume $u$ is continuous inside the domain and bounded near the boundary. $\endgroup$ Aug 23, 2019 at 9:49
  • $\begingroup$ If I am not wrong, I obtain that there exist a constant $C$ which depends on $\Omega$ whenever $\Omega$ is a unit ball, which is different from the Laplacian case. $\endgroup$
    – GabS
    Aug 24, 2019 at 9:06
  • $\begingroup$ You might create the tag "fractional-laplacian", but "fractional" is a too imprecise tag. $\endgroup$
    – YCor
    Feb 16, 2021 at 12:16

1 Answer 1

1
$\begingroup$

If $u$ is a solution of the above problem, then $u$ is said to be harmonic in $\Omega$ with respect to $(-\Delta)^s$. If $u$ is continuous and bounded in $\Omega$, and $\Omega$ satisfies the exterior cone condition, then $$ u(x) = \int_{\mathbb{R}^N \setminus \Omega} g(y) P_\Omega(x, dy) ,$$ where $P_\Omega(x, \cdot)$ is a probability measure on the complement of $\Omega$. This measure is called the harmonic measure for $(-\Delta)^s$, and it has a nice probabilistic interpretation: it is the distribution of the isotropic $2s$-stable Lévy process started at $x$ and stopped at the time of first exit from $\Omega$. As a consequence, $|u(x)| \leqslant \|g\|_\infty$, as desired.

You can find the above result in Bogdan's 1997 paper [1], see Lemma 17 therein. For further discussion and other references, you may like to take a look at my survey paper [2], in particular — Theorem 7.2 therein.

[1] K. Bogdan, The boundary Harnack principle for the fractional Laplacian, Stud. Math., 123(1) (1997), 43–80, available at http://matwbn.icm.edu.pl/ksiazki/sm/sm123/sm12313.pdf.

[2] M. Kwaśnicki, Fractional Laplace Operator and its Properties, in: A. Kochubei, Y. Luchko, Handbook of Fractional Calculus with Applications. Volume 1: Basic Theory, De Gruyter Reference, De Gruyter, Berlin, 2019, DOI:10.1515/9783110571622-007.

$\endgroup$
0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.