On exponential densities and limit ratios of subsets of \(\mathbb{N}\) (Q779951)

Let \(A=\{a_1,a_2,\dots\},\) with \(a_1<a_2<\cdots,\) be an infinite subset of the set \({\mathbb N}\) of positive integers. The lower and upper \textit{limit ratios} of \(A\) are defined by \[\underline\rho(A)=\liminf_{n\rightarrow+\infty}{a_n/a_{n+1}},\] \[\overline\rho(A)=\limsup_{n\rightarrow+\infty}{a_n/a_{n+1}}.\] The lower and upper \textit{exponential densities} of \(A\) are defined by \[\underline\varepsilon(A)=\liminf_{n\rightarrow+\infty}{\log|A\cap[1,n]|/{\log n}},\] \[\overline\varepsilon(A)=\limsup_{n\rightarrow+\infty}{\log|A\cap[1,n]|/{\log n}}.\] In the paper [\textit{G. Grekos} et al., Unif. Distrib. Theory 6, No. 2, 117--130 (2011; Zbl 1313.11027)] it is proved that if \(\overline\rho(A) <1,\) then \(\overline\varepsilon(A)=\underline\varepsilon(A)=0.\) In the paper under review the authors prove that if \(\overline\rho(A) =1,\) then all is possible for the three other quantities. Precisely, they prove the next theorem: ``Given three real numbers \(\alpha,\beta,\gamma\) belonging to [0,1] with \(\alpha\le\beta,\) then there exists an infinite subset \(A\) of \({\mathbb N}\) such that \[\underline\varepsilon(A)=\alpha, \overline\varepsilon(A)=\beta, \underline\rho(A)=\gamma, \overline\rho(A)=1."\] The proof goes through two distinct constructions, a first one for \(\beta=0\) and a second one when \(\beta>O.\) The paper contains useful information concerning the use of the notions in the title (exponential densities, limit ratios) in various area of mathematics: number theory, fractal geometry, theoretical computer science.