Arithmetic properties of numbers with restricted digits (Q4811230)

Let \(g>1\) be an integer and \({\mathcal D}\subset \{0,1,\dots,g-1\}\) be a subset of digits in base \(g\). For a positive integer \(n\) put \({\mathcal A}_g(n,{\mathcal D})\) for the set of all positive integers \(s<g^n\) which are coprime to \(g\) and whose base \(g\) digits are in \({\mathcal D}\). In the paper under review, the authors investigate various arithmetic properties of the members of \({\mathcal A}_g(n,{\mathcal D})\). For example, they compute the sum of the values of the Euler function of \(s\) and the sum of divisors function of \(s\) when \(s\) ranges over \({\mathcal A}_g(n,{\mathcal D})\). They investigate the number of elements of \({\mathcal A}_g(n,{\mathcal D})\) that are relatively prime to a given number \(t\geq 1\), the number of elements of \({\mathcal A}_g(n,{\mathcal D})\) that are \(\ell\)-powerfree, as well as the number of pairs of relatively prime elements in \({\mathcal A}_g(n,{\mathcal D})\). As expected, the above quantities are close to their expected values for a wide range in the values of \(\#{\mathcal D},~k\) and \(n\). Similar questions are investigated for the set \({\mathcal B}(n,k)\) consisting of those positive integers \(s<2^n\) having precisely \(k\) ones in their binary representation. The proofs are elementary (inclusion-exclusion principle and changes in orders of summations) except that they use a result from [\textit{S. Konyagin}, Period. Math. Hung. 42, 145--162 (2001; Zbl 0980.11006)]. One of the results from this paper gives a partial answer to an open question from [\textit{P. Erdős}, \textit{C. Mauduit} and \textit{A. Sárközy}, J. Number Theory 70, 99--120 (1998; Zbl 0923.11024)].