The stationary critical points of the fractional heat flow

TL;DR

This paper investigates the critical points of solutions to the fractional heat equation, establishing conditions for symmetry and shape of the domain based on the behavior of solutions, extending classical results to the fractional, nonlocal setting.

Contribution

It extends classical symmetry and shape results for the heat equation to the fractional case, accounting for nonlocal effects and providing new insights into domain characterization.

Findings

Origin's critical point condition linked to initial data symmetry.

Domain symmetry characterized by critical point behavior.

New methods applicable to classical and fractional heat flows.

Abstract

We study the spatial critical points of the solutions $u=u(x,t)$ of the fractional heat equation. For the Cauchy problem, we show that the origin $0$ satisfies $\nabla_x u(0,t) = 0$ for $t>0$ if and only if the initial data satisfy a balance law of the form $\int_{\mathbb{S}^{N-1}} \omega u_0(r\omega) \, \mathrm{d} \omega=0$ for a. e. $r \ge 0$. Moreover, for the Dirichlet initial-boundary value problem, we prove two symmetry results: $\Omega$ is a ball centered at the origin if and only if $\nabla_x u(0,t) = 0$ for $t>0$ provided that the initial data satisfies the above-mentioned balance law; $\Omega$ is centrosymmetric if and only if $\nabla_x u(0,t) = 0$ for $t>0$ provided that the initial data is centrosymmetric. These results extend some theorems obtained by Magnanini and Sakaguchi in 1997-1999 for the (local) heat equation to the fractional context. These extensions are nontrivial because of the nonlocal nature of the fractional Laplacian. Among others, the proof of the characterization of a ball in the Dirichlet initial-boundary value problem for the fractional heat flow not only works for the classical heat flow but also gives a new insight into the problem.