Metric discrepancy results for geometric progressions perturbed by irrational rotations (Q2204098)

From MaRDI portal
Revision as of 22:53, 25 July 2024 by Daniel (talk | contribs) (‎Created claim: Wikidata QID (P12): Q127024899, #quickstatements; #temporary_batch_1721940658196)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
scientific article
Language Label Description Also known as
English
Metric discrepancy results for geometric progressions perturbed by irrational rotations
scientific article

    Statements

    Metric discrepancy results for geometric progressions perturbed by irrational rotations (English)
    0 references
    0 references
    2 October 2020
    0 references
    For every \(\theta\in(-\infty,-1)\cup(1,\infty)\) and for a.e. \(x\), it is known that the sequence \(\{\theta^kx\}\) is uniformly distributed modulo \(1\). The speed of convergence sensitively depends on the algebraic nature of \(\theta\). The following law of iterated logarithm was proved in \textit{K. Fukuyama} [Acta Math. Hung. 118, No. 1--2, 155--170 (2008; Zbl 1241.11090)], \textit{K. Fukuyama} [Monatsh. Math. 171, No. 1, 33--63 (2013; Zbl 1334.11060)]: There are constants \(\Sigma_\theta\) and \(\Sigma_\theta^\ast\) such that \[ \limsup_{N\to\infty}\frac{ND_N\{\theta^k x\}}{\sqrt{2N\log\log N}}=\Sigma_\theta,\ \limsup_{N\to\infty}\frac{ND_N^\ast\{\theta^k x\}}{\sqrt{2N\log\log N}}=\Sigma_\theta^\ast \ \ \text{a.e.} \] Here \(D_N\) denotes the discrepancy \[ D_N\{x_k\}:=\sup_{0\leq a<b\leq 1}\left\vert \frac{1}{N}\sum_{k=1}^N{1}_{[a,b)}(\langle x_k\rangle)-(b-a)\right\vert,\] where \(\langle x\rangle \) is the fractional part of \(x\). To define \(D_N^\ast,\) one sets \(a=0\) in the definition of \(D_N\). \par The main result of the paper is as follows: Given any \(\vert \theta\vert >1\) and \(\gamma\not\in\mathbb{Q}\). If in the limits above one replaces \(\{\theta^k x\}\) with the perturbed sequence \(\{\theta^k x+ k\gamma\}\), then both limits become \(1/2\). Moreover, if \(\theta^r\not\in\mathbb{Q},r\in\mathbb{N}\), then this is true for every \(\gamma\in\mathbb{R}\). This speed of convergence is identical with that of the sequence of uniformly distributed independent random variables.
    0 references
    0 references
    0 references
    0 references
    0 references
    discrepancy
    0 references
    lacunary sequence
    0 references
    law of the iterated logarithm
    0 references
    0 references
    0 references