Random sum-free subsets of abelian groups (Q2017135)

Let \(G\) be an abelian group. A nonempty subset \(A\) of \(G\) is called sum-free if \(A\) does not contain \(x + y\), where \(x,y\in A\). The \(p\)-random subset of a set is that random subset of it where each element is chosen independently of all other elements with probability \(p\). Let \(G_n\) be a sequence of abelian groups with \(|G_n| = n\). A function \(p_c(n)\) is called threshold function for a property \(\mathcal{P}\) if the \(p\)-random subset \(B \subseteq G_n\) has \(\mathcal{P}\) with high probability when \(p \gg p_c(n)\) and \(B\) does not have \(\mathcal{P}\) with high probability when \(p \ll p_c(n)\). A threshold function is called sharp if moreover for every \(\varepsilon > 0\) the above holds with \(p > (1+\varepsilon)p_c(n)\) and \(p < (1-\varepsilon)p_c(n)\). Consider the group \(G = \mathbb{Z}_{2n}\) and let \(O_{2n}\) be the set of odd numbers. For a subset \(B \subseteq G\) let \(SF_{0}(B)\) be the collection of maximum-size sum-free subsets of \(B\). In this paper the authors give a sharp threshold for the property that \(SF_{0}(G_p) = \{G_p \cap O_{2n}\}\), where \(G_p\) is a \(p\)-random subset of \(\mathbb{Z}_{2n}\). They determine the threshold for all abelian groups the order of which has a prime divisor of the form \(3k+2\) as well. They also show that for certain groups the threshold can be much larger. In particular, for the group \(\mathbb{Z}_n\), where \(n = 3q\), where \(q\) is a prime of the form \(3k+1\) and when \(n\) is a prime of the form \(3k+2\). They also prove a similar result for the group \(\mathbb{Z}_2^k\). The proof is based on the applications of some probabilistic tools and extremal results on abelian groups. They used a lower bound due to \textit{B. Green} and \textit{I. Z. Ruzsa} [Isr. J. Math. 147, 157--188 (2005; Zbl 1158.11311)] and a lower bound due to \textit{V. F. Lev} et al. [Isr. J. Math. 125, 347--367 (2001; Zbl 1055.20043)] for the cardinality of sum-free subsets in certain abelian groups. On the other hand they applied some well known probabilistic inequalities such as Chernoff inequality, Janson inequality and the FKG inequality. A key tool in the proofs is the transference theorem of \textit{D. Conlon} and \textit{W. T. Gowers} [Ann. Math. (2) 184, No. 2, 367--454 (2016; Zbl 1351.05204)].