arXiv Analytics

Sign in

arXiv:math/0407318 [math.PR]AbstractReferencesReviewsResources

α-Continuity Properties of Stable Processes

R. D. DeBlassie, Pedro J. Mendez-Hernandez

Published 2004-07-19Version 1

Let $D$ be a domain of finite Lebesgue measure in $\bR^d$ and let $X^D_t$ be the symmetric $\alpha$-stable process killed upon exiting $D$. Each element of the set $\{\lambda_i^\alpha\}_{i=1}^\infty$ of eigenvalues associated to $X^D_t$, regarded as a function of $\alpha\in(0,2)$, is right continuous. In addition, if $D$ is Lipschitz and bounded, then each $ \lambda_i^\alpha$ is continuous in $\alpha$ and the set of associated eigenfunctions is precompact. We also prove that if $D$ is a domain of finite Lebesgue measure, then for all $0<\alpha<\beta\leq 2$ and $i\geq 1$, \[\lambda_i^\alpha \leq [ \lambda^\beta_i]^{\alpha/\beta}.\] Previously, this bound had been known only for $\beta=2$ and $\alpha$ rational.

Related articles: Most relevant | Search more
arXiv:math/0211192 [math.PR] (Published 2002-11-12, updated 2002-11-18)
Concentration of norms and eigenvalues of random matrices
arXiv:0801.0330 [math.PR] (Published 2008-01-02)
Properties of Expectations of Functions of Martingale Diffusions
arXiv:math/0606624 [math.PR] (Published 2006-06-24, updated 2007-12-12)
Eigenvalues of Euclidean Random Matrices