On additive and multiplicative decompositions of sets of integers with restricted prime factors. II: Smooth numbers and generalizations (Q2040149)

This paper is a natural continuation of the paper [\textit{K. Győry} et al., Indag. Math., New Ser. 32, No. 2, 365--374 (2021; Zbl 1470.11258)]. We start giving some definitions needed to state the main results. Let \(\mathcal{A}\), \(\mathcal{B}\) and \(\mathcal{C}\) be infinite sets of non-negative integers. Then, we write \(\mathcal{B}+\mathcal{C}\) for the Minkowski sum of the sets, and \(\mathcal{B}\cdot \mathcal{C}=\{b\cdot c: b\in \mathcal{B}, c\in \mathcal{C}\}\). If \(|\mathcal{B}|, |\mathcal{C}|\geq 2\) we say that \(\mathcal{A}\) is a-reducible if \(\mathcal{A}=\mathcal{B}+\mathcal{C}\). Otherwise, if there does not exist such sets \(\mathcal{B}\) and \(\mathcal{C}\) we say that \(\mathcal{A}\) is a-irreducible. A similar definition is done changing the sum by the product (and changing letter `a' by `m' in the notion of a-reducible and a-irreducible). We need a couple more definitions. Given to infinite sets \(\mathcal{A}\) and \(\mathcal{A}'\), we say that they are asymptotically equal if they coincide from a point on (We denote that by \(\mathcal{A} \sim \mathcal{A}'\)). Finally, we say that \(\mathcal{A}\) is totally a-irreducible if for all \(\mathcal{A}'\sim \mathcal{A}\), then \(\mathcal{A}'\) is a-irreducible. Giving now a number \(n\), we denote by \(p^{+}(n)\) its greatest prime factor. For a (typically growing) function \(y(n)\) on \(n\) we say that a number is \(y\)-smooth if \(p^{+}(n)\leq y(n)\). Then, given an infinite family \(\mathcal{F}\), we denote by \(\mathcal{F}_y\) the set of \(y\)-smooth numbers in \(\mathcal{F}\). The results in [loc. cit.] deal with the case where \(y(n)\) grows slowly. More precisely: if \(y(n)\leq C \log (n)\) for an small and concrete constant \(C\), then \(\mathcal{F}_y\) is totally a-irreducible. A similar result is proven for \(\mathcal{G}_{y}:=\mathcal{F}_{y}+\{1\}\) and concerning totally m-irreducibility. The main contribution of the present paper is to continue the investigation when \(y(n)\) grows like a constant in \(n\). In this regime it is not possible to assure totally \(a\)-irreducibility. More precisely, the authors are able to show that if \(y(n)\) is monotone function such that \(n/2<y(n)<n\), then \(\mathcal{F}_y\) is NOT totally a-irreducible. Related results on this regime for \(y\) are also proven (see Theorem 1.2 and 1.3), as well as extensions to \(k\)-term sums of elements on a given set.