Universal constraints on the location of extrema of eigenfunctions of non-local Schrödinger operators (Q1737537)

From MaRDI portal





scientific article
Language Label Description Also known as
default for all languages
No label defined
    English
    Universal constraints on the location of extrema of eigenfunctions of non-local Schrödinger operators
    scientific article

      Statements

      Universal constraints on the location of extrema of eigenfunctions of non-local Schrödinger operators (English)
      0 references
      0 references
      0 references
      23 April 2019
      0 references
      It is well-known that for a Bernstein function $\Psi$, [\textit{R. L. Schilling} et al., Bernstein functions. Theory and applications. 2nd revised and extended ed. Berlin: de Gruyter (2012; Zbl 1257.33001)], the operator $\Psi(-\Delta)$ is the generator of a Lévy process which can be obtained from a Brownian motion by an independent time-change (subordination in the sense of Bochner); $-\Psi(-\Delta)$ can be interpreted in any reasonable functional-calculus sense. The authors perturb $-\Psi(-D)$ by a potential $V$ and consider the Dirichlet eigenvalue problem for the operator $H=\psi(-\Delta)+V$ on a bounded convex domain $D\subset\mathbb R^d$: $Hu = \lambda u$ on $D$ and $u=0$ on $D^c$. If $\Psi$ grows fast enough as $|x|\to\infty$, the operator $\Psi(-\Delta)$ has, as operator in $L^2(D)$, countably many non-negative eigenvalues $\lambda_i^D$, and this property is inherited by the perturbed operator $H$; the eigenvalues are $\lambda_i^{D,V}$. In addition, it is assumed that the Bernstein function $\Psi$ satisfies a lower scaling condition: There exist $\mu>0$ and $c\in (0,1]$ such that $\Psi(\gamma x)\geq c\gamma^\mu\Psi(x)$ holds for all $x>0$ and $\gamma\geq 1$. The potential $V$ is assumed to be a $\Psi$-Kato class potential, if the negative part $V^-$ and, for each compact set, the localized positive part $V^+ 1_{K}$ satisfy the condition \[ \lim_{t\to 0}\sup_{x\in\mathbb R^d}\mathbb{E}^x\int_0^t |f(X_s)|\,ds =0,\quad f\in\{V^-,V^+ 1_K\}. \] The main result of the paper states the following: Let $V$ be a $\Psi$-Kato class potential with $V^-\in L^p$ and $p>d/(2\mu)$, let $\phi$ be a non-zero eigenfunction for the eigenvalue $\lambda^{D,V}$ such that $|\phi|$ has a global maximum at $x^*\in D$, $r=\text{dist}(x^*,\partial D)$. Then there are constants $\Theta_1, \Theta_2>0$ such that \[ \Theta_1\|V^-\|_{L^p}^{1/\eta}-\inf_D V^+ + \lambda^{V,D}\geq\Theta_2\Psi(r^{-2}); \] here, $\eta = 1- d/(2\mu p)$. If $V$ is bounded this estimate is improved to \[ \sup_D V^- - \inf_D V^+ + \lambda^{D,V} \geq \Theta_3\Psi(r^{-2}). \] Some applications are given, for example a Faber-Krahn type inequality.
      0 references
      non-local Schrödinger operators
      0 references
      subordinate Brownian motion
      0 references
      Dirichlet exterior value problem
      0 references
      principal eigenvalues and eigenfunctions
      0 references
      hot spots
      0 references
      Faber-Krahn inequality
      0 references
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references