Small-sum pairs in abelian groups (Q719132): Difference between revisions

In [``Kneser's theorem for upper Banach density'', J. Théor. Nombres Bordx. 18, No. 2, 323--343 (2006; Zbl 1138.11044)], \textit{P. Bihani} and \textit{R. Jin} asked the following question: Let \(A\), \(B\) be subsets in an abelian group \(G\) such that \(|A|=|B|=k\) and \(|A+A|=|A+B|=2k-1\). Are always \(A\) and \(B\) translates of each other? The goal of this paper is to answer this question in the affirmative under appropriate hypotheses and to provide an explicit description of the counterexamples. If for two subsets \(X\), \(Y\) of \(G\) and \(c\in G\) we denote \(\upsilon _{c}(X,Y) \) the number of pairs \((x,y)\in X\times Y\) such that \(c=x+y\), it is proved that an element \(g\in G\) exists such that \(A=g+B\) if \(A+B\) is aperiodic or if elements \(a\in A\), \(\in B\) exist such that \(\upsilon _{a+b}(A,B)=\upsilon _{a+a}(A,A)=1\). The results heavily rely on on the description of the so-called critical pairs (of subsets) given by \textit{J. H. B. Kemperman} [Acta Math. 103, 63--88 (1960; Zbl 0108.25704)].