Inverse results for restricted sumsets in \(\mathbb{Z}/p\mathbb{Z}\) (Q6571320)

Let \(p\) be a prime, \(A\) and \(B\) be subsets of \(\mathbb{Z}/p\mathbb{Z}\) and \(S\) be a subset of \(A \times B\). Let \(A +^S B=\{a+b:(a,b)\in S\}\). This kind of subset is called a restricted sumset. There are classic results in additive number theory for unrestricted sumsets. One of the main motivations for working with restricted sumsets is to generalize those classical results or at least find similar results. An important result in \(\mathbb{Z}\) for unrestricted sumsets is Freiman's \(3k -4\) theorem, see \textit{D. J. Grynkiewicz} [Structural additive theory. Based on courses given at Karl-Franzens-Universität Graz, Austria, 2008--2012. Cham: Springer (2013; Zbl 1368.11109)], \textit{M. Huicochea} [Unif. Distrib. Theory 14, No. 2, 43--68 (2019; Zbl 1477.11019)], \textit{M. B. Nathanson} [Additive number theory. Inverse problems and the geometry of sumsets. New York, NY: Springer (1996; Zbl 0859.11003)], \textit{T. Tao} and \textit{V. H. Vu} [Additive combinatorics. Cambridge: Cambridge University Press (2006; Zbl 1127.11002)]. The main motivation of this paper is to get a nontrivial inverse result for restricted sumsets in \(\mathbb{Z}/p\mathbb{Z}\). Also, these inverse results will have applications for unrestricted sumsets. One of the most important inverse results for restricted sumsets is the Balog-Szemerédi-Gowers theorem by \textit{V. Vu}. However, when the restricted sumset \( A +^S {+}B\) is small and \(S\) is almost \(A\times B\), sharper results can be found. The first result of this paper is a result of this kind. In the first inverse result of this paper, it is shown that if \( \Big\vert A+^S B \Big\vert \) and \(\vert (A\times B)\setminus S \vert \) are small, then \(A\) has a big subset with small difference set. In the second theorem of this paper, it is shown that if \( A+^S B, \vert A\vert \) and \(\vert B\vert \) are small, then big parts of \(A \) and \( B\) are contained in short arithmetic progressions with the same difference. As an application of this result, an inverse of Pollard's theorem is obtained.