Generalized Sidon sets (Q1959677): Difference between revisions

Let \(A\) be a subset of \(\mathbb Z\), the set of integers. Its \textit{representation function} is defined as follows: for every \(x\in{\mathbb Z},\) put \[ r_A(x)=\#\{(a,a')\in A^2~;~a+a'=x\}. \] \(A\) is called a \(g\)-\textit{Sidon set} if \(r_A(x)\leq g\) for every \(x\in{\mathbb Z}\). For each positive integer \(n,\) define \[ \beta_g(n)=\max\{\#A~;~A\subset\{1,\dots,n\},A~\text{is a \(g\)-Sidon set}\}. \] Many questions concerning \[ {\overline\beta}_g=\limsup_{n\rightarrow+\infty}\beta_g(n)/{\sqrt n}~\text{and}~{\underline\beta}_g=\liminf_{n\rightarrow+\infty}\beta_g(n)/{\sqrt n} \] are open, especially for \(g\geq4.\) The main result of the paper under review is that \[ \lim_{g\rightarrow+\infty}{{{\overline\beta}_g}\over{\sqrt g}}= \lim_{g\rightarrow+\infty}{{{\underline\beta}_g}\over{\sqrt g}}=\sigma \] where \(\sigma\) is a constant, appeared in [\textit{A. Schinzel} and \textit{W. M. Schmidt}, Acta Arith. 104, No. 3, 283--296 (2002; Zbl 1017.11013)] and defined in the folowing way: ``Consider all nonnegative real functions \(f\) satisfying \(f(x)=0\) for \(x\not\in[0,1],\) and \[ \int_0^1 f(t) f(x-t)~ dt~\leq 1 \] for all \(x.\) Define the constant \(\sigma\) by \[ \sigma=\sup\int_0^1 f(x)~dx \] where the supremum is taken over all functions \(f\) satisfying the above restrictions.'' The value of \(\sigma\) is not known. For estimations see [\textit{M. Matolcsi} and \textit{C. Vinuesa}, J. Math. Anal. Appl. 372, No. 2, 439--447 (2010; Zbl 1247.11031)]. In the proofs the authors use modular reductions. This gives rise to further interesting investigations concerning \(g\)-Sidon sets in \({\mathbb Z}/q{\mathbb Z}\) and in \(({\mathbb Z}/p{\mathbb Z})^2,\) with \(p\) prime.