Iterated sumsets and subsequence sums (Q1671777): Difference between revisions

Let \(G\) be an abelian group (finite or not) and let \(A\subseteq G\) be a finite set such that the difference set \(A-A = \{a_1-a_2 : a_1,a_2 \in A\}\) generates \(G\). Also, let \(n\ge 3\) be an integer, the sumset \(nA = \{ a_1+\dots+a_n : a_1,\dots, a_n \in A \}\) has zero stabilizer and \[ |nA| < (|A|+1)n - 3 \,.\tag{1} \] The paper is devoted to study the structure of sets satisfying (1) and the author gives a complete description of such sets. After that he applies this inverse result to generalize a theorem of \textit{J. E. Olson} [J. Number Theory 9, 63--70 (1977; Zbl 0351.20032)] which is in turn a refinement of the famous Erdős-Ginzburg-Ziv theorem [\textit{P. Erdős} et al., Bull. Res. Council Israel 10F, 41--43 (1961; Zbl 0063.00009)], in several directions. For example, consider a sequence \(S\) of some elements of \(G\), \(|S|\ge n+|G|-1\), where \(1\le n\) be the maximal multiplicity in the sequence \(S\). Suppose that for every proper subgroup \(H\) and \(\alpha \in G\) there are at least \(|G/H|-1\) terms of \(S\) lying outside the coset \(H+\alpha\). If \(n\ge \exp (G)\), then any element of \(G\) can be expressed as a sum of several elements of \(S\).