scientific article; zbMATH DE number 3918485

From MaRDI portal
Publication:3695347





If \(A\) is an infinite subset of the set \({\mathbb{N}}\) of natural numbers, \(A(x)\) denotes the number of elements of \(A\) which are \(\leq x\). The main theorem states: If \(k\) and \(n\) are given natural numbers \(>1\) and if \(A(x) \geq \epsilon \cdot x^{1-1/n}\) for some positive \(\epsilon\) and all \(x\) of a final segment of \({\mathbb{N}}\), then there exist \(k\) disjoint \(n\)-element subsets of A which are congruent by translation. Of course, this also implies that \(n\) disjoint \(k\)-element subsets of \(A\) exist which are congruent by translation. This improves an earlier result of the first author for \(k = n = 2\) on \(B_2\)-sequences, which was published in a paper of \textit{A.Stöhr} [J. Reine Angew. Math. 194, 111-140 (1955; Zbl 0066.03101)]. One obtains the corollary that for every two natural numbers \(k,n\) the set of prime numbers has \(k\) disjoint \(n\)-element subsets which are congruent by translation. Concerning the sharpness of the theorem there holds: If \(0 < \alpha < 1-1/k-1/n+1/kn\) then for all sufficiently large natural numbers m there exists a subset of \(\{1,...,m\}\) which has at least \(m^{\alpha}\) elements but no \(k\) disjoint congruent \(n\)-element subsets.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3695347)