On general densities and intersectivity (Q648957)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | On general densities and intersectivity |
scientific article |
Statements
On general densities and intersectivity (English)
0 references
29 November 2011
0 references
The author generalizes for higher dimensions a theorem of \textit{I. Z. Ruzsa} [On difference sets, Stud. Sci. Math. Hung. 13, 319--326 (1978; Zbl 0423.10027)] proved for dimension 1. Let \(E\) be a subset of \({\mathbb N}^d\). Its [\textit{upper}] \textit{Banach density} \(B(E)\) is defined as follows. Let \(R=(R_k)_{k\geq1}\) be a sequence of rectangles in \({\mathbb N}^d\) with sides parallel to the axes. Then put \[ B(E)=\sup_R~(\limsup_{k\rightarrow\infty}{{|E\cap R_k|}\over {|R_k|}}) \] where the supremum is taken over all such sequences \(R\). The author defines the \textit{goodness} of a sequence \(R\) of rectangles. Roughly, the sequence \(R\) has this property if the sequence \((n_k)\) of the corresponding ``south-west'' corners \(n_k\) of \(R_k\), is rapidly increasing but the sequences of the lengths of edges increase not so rapidly. The precise technical condition for goodness is not presented in this review; here are two examples. Let \((\ell_{1,k},\dots,\ell_{d,k})\) denote the sides of the rectangle \(R_k\). Suppose that \(\ell_{i,k}\leq\ell_{i,k+1}\) for all \(i\), \(1\leq i\leq d\), and for all \(k.\) -- Example 1.- If for all \(k,\) \(n_k=4^k\) and \(\min\{\ell_{i,k}~; 1\leq i\leq d\}\geq 2^k,\) then the sequence \(R\) is not good. -- Example 2.- If for all \(k,\) \(n_k=2^{2^k}\) and \(\max\{\ell_{i,k}~;1\leq i\leq d\}\leq\sqrt{n_k},\) then the sequence \(R\) is good. The author proves the following theorem. Let \(E\subseteq{\mathbb N}^d\) with \(B(E)>0\). Let \(R\) be a good sequence of rectangles. Then there exists a subset \(S\) of \({\mathbb N}^d\) such that: 1) \(S\) is well distributed in the rectangles \(R_k\) in the sense that \[ \lim_{k\to +\infty}{{|S\cap R_k|}\over{|R_k|}} \] exists and is greater than or equal to \(B(E)\). 2) \(S\) satisfies the following intersectivity property: for every finite subset \(\{m_1,\dots,m_r\}\) of \(S\) we have \[ B(E\cap(E+m_1)\cap\dots\cap(E+m_r))>0. \] In the proof a lemma proved in [\textit{R. L. Jones} and \textit{J. Olsen}, ``Multi-parameter moving averages'', Almost every where convergence II. Proc. 2nd Int. Conf., Evanston/IL (USA) 1989, 127--149 (1991; Zbl 0752.28008)], is used.
0 references
densities
0 references
ergodic theorems
0 references
intersectivity
0 references
0 references