Search arXiv⌕ Search
Home/Archive/arXiv

arXiv · math/0412260

Estimates and identities for the average distortion of a linear transformation

Igor Rivin
View original source ↗Read source document ↗Export metadata ↓

Abstract

For a linear transformation A from Rn to Rn, we give sharp bounds for the average distortion of A, that is, the average value of log of the euclidean norm of Au over all unit vectors u. This is closely related to the results of the author's paper math.DS/0312048.

Explore related subjects

math.CAmath.DS

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Igor Rivin. 2004-12-13. Estimates and identities for the average distortion of a linear transformation. https://arxiv.org/abs/math/0412260

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Exponential decay of threshold correlations for symmetric periodic functions

We study how often $n$ distinct integer dilates of a periodic function can all lie above a fixed threshold or all lie below its negative. We consider continuous functions $f:\mathbb R\to[-1,1]$ that are $1$-periodic, even, strictly decreasing on $[0,1/2]$, and satisfy $f(0)=1$ and $f(t+1/2)=-f(t)$. This class includes the cosine and the triangle wave. For $0\leqα<1$, let $ρ_α\in(0,1/4]$ be determined by $f(ρ_α)=α$, and let $F_α(n)$ denote the infimum of these proportions over all sets of $n$ distinct positive integer frequencies. We prove that $F_α(n)\geq 4ρ_α^n$ and that the limit $\lim_{n\to\infty}F_α(n)^{1/n}$ exists for every threshold. When $ρ_α^{-1}$ is an integer, we obtain \begin{align} F_α(n) &\leq C_αn^{\log_2((1-ρ_α)/(1-2ρ_α))}ρ_α^n, \end{align} and hence the limit equals $ρ_α$. Otherwise, the limit belongs to the interval $[ρ_α,\lfloorρ_α^{-1}\rfloor^{-1}]$. The upper bound in the reciprocal case uses a cyclic covering construction previously studied in the context of circular graphs. We give a self-contained proof of the corresponding optimal modulus.

math.CA↗
arXiv2026

Endpoint estimates for spherical maximal operators acting on radial functions in hyperbolic spaces

In this article we investigate spherical maximal operators $\mathfrak m^α$ of real order $α$ acting on radial functions in the hyperbolic space $\mathbb H^n$. We show that for every $n\geq 2$, $1 (1-n)/2$, $\mathfrak m^α$ is bounded on $L^p_{\mathrm{rad}}(\mathbb H^n)$ if and only if \begin{align*} \begin{cases} α>1-n+\frac np, & 1 \frac1p-\frac{n-1}2, & 2 \frac{2n}{n-1}. \end{cases} \end{align*} Moreover, we show that for $1<p\leq 2n/(n-1)$ and \begin{align*} α= \begin{cases} 1-n+\frac np, & 1<p\leq 2, \frac1p-\frac{n-1}2, & 2<p\leq\frac{2n}{n-1}, \end{cases} \end{align*} $\mathfrak m^α$ is bounded from $L^{p,1}_{\mathrm{rad}}(\mathbb H^n)$ to $L^{p,\infty}_{\mathrm{rad}}(\mathbb H^n)$.

math.CA↗
arXiv2026

On $q$-middle convolution and generalized $q$-hypergeometric equation

We investigate higher-order $q$-difference equations obtained by successive applications of a $q$-analogue of addition and the $q$-middle convolution. We construct an $n$-dimensional system of linear $q$-difference equations and derive a single $q$-difference equation whose coefficients are polynomials of degree at most $1$ in the independent variable. Under certain conditions on the parameters, the resulting equation is a generalized $q$-hypergeometric equation of order $n$. We also obtain a multiple Jackson integral representation of solutions and give sufficient conditions for its convergence. By suitably choosing the functions appearing in the integrand and specializing the parameters, we obtain $q$-hypergeometric series solutions of type ${}_n ϕ_{n-1}$.

math.CA↗