Large deviations for the maximum and reversed order statistics of Weibull-like variables

Fuente: arXiv
Saved in:
Bibliographic Details
Main Author: Jansen, Sabine
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916262365364224
author Jansen, Sabine
author_facet Jansen, Sabine
contents Motivated by metastability in the zero-range process, we consider i.i.d.\ random variables with values in $\N_0$ and Weibull-like (stretched exponential) law $\mathbb P(X_i =k) = c \exp( - k^α)$, $α\in (0,1)$. We condition on large values of the sum $S_n= μn + s n^γ$ and prove large deviation principles for the rescaled maximum $M_n /n^γ$ and for the reversed order statistics. The scale is $n^γ$ with $γ= 1/(2-α)$; on that scale, the big-jump principle for heavy-tailed variables and a naive normal approximation for moderate deviations yield bounds of the same order $n^{γα} = n^{2γ-1}$, the speed of the large deviation principles. The rate function for $M_n/n^γ$ is non-convex and solves a recursive equation similar to a Bellman equation.
format Preprint
id arxiv_https___arxiv_org_abs_2405_17319
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Large deviations for the maximum and reversed order statistics of Weibull-like variables
Jansen, Sabine
Probability
60F10, 60G50, 60K35
Motivated by metastability in the zero-range process, we consider i.i.d.\ random variables with values in $\N_0$ and Weibull-like (stretched exponential) law $\mathbb P(X_i =k) = c \exp( - k^α)$, $α\in (0,1)$. We condition on large values of the sum $S_n= μn + s n^γ$ and prove large deviation principles for the rescaled maximum $M_n /n^γ$ and for the reversed order statistics. The scale is $n^γ$ with $γ= 1/(2-α)$; on that scale, the big-jump principle for heavy-tailed variables and a naive normal approximation for moderate deviations yield bounds of the same order $n^{γα} = n^{2γ-1}$, the speed of the large deviation principles. The rate function for $M_n/n^γ$ is non-convex and solves a recursive equation similar to a Bellman equation.
title Large deviations for the maximum and reversed order statistics of Weibull-like variables
topic Probability
60F10, 60G50, 60K35
url https://arxiv.org/abs/2405.17319