An analytic approach to cardinalities of sumsets (Q2151183): Difference between revisions

The paper under review studies what properties of a subset of \(\mathbb{Z}^d\) make its sumset large. For a finite set \(U\subset \mathbb{Z}^d\), the \textit{additive tripling} of \(U\) is \(\beta(U)= \inf_{A,B} \vert A+B+U\vert /\sqrt{\vert A\vert \vert B\vert}\). The paper defines recursively \textit{quasicubes}, a generalization of the hypercubes. In \(\mathbb{Z}^d\), quasicubes have \(2^d\) elements. Among many related results, the paper shows that subsets of quasicubes have maximal \(\beta\). As a corollary, for a \(d\)-dimensional quasicube \(U\), and any \(V\subset U\) and \(k\geq 2\), the paper obtains \(\vert (2^k-1)V\vert \geq \vert V\vert ^k\). The proof techniques include a discrete Prékopa-Leindler inequality for quasicubes, a discrete Brunn-Minkowski inequality for quasicubes, tensorisation, etc. Typically, the results of the paper can be formulated in a more general setting, namely in torsion-free commutative groups, not just in \(\mathbb{Z}^d\).