Pair correlation of fractional parts derived from rational valued sequences. II (Q495280)

The pair correlation function can be used to measure the distribution of spacings between different elements of infinite sequences and provides a way to compare the behavior of sequences. The authors define a \textit{rational valued vector sequence \(\vec x\)}, for some fixed \(k\) and \(r\in \mathbb N\) as a map \(\vec x:\mathbb N^k\rightarrow\mathbb Q^r\). Complementing the results of their previous paper [ibid. 151, 147--158 (2015; Zbl 1323.11059)], the authors study the pair correlation function for the fractional parts of sequences \(\vec t\cdot\vec x\), where \(\vec x\) is a rational valued vector sequence and \(\vec t\in\mathbb R^r\). The main result of the paper is a new class of sequences \(\vec x\) whose pair correlation function behaves like that of random sequences for almost all real vectors \(\vec t\); i.e., the particular class of sequences \[ \vec{x}(m,n)=\left(\frac{p_1^m p_2^n}{\vec{b}(m,n)},\frac{p_3^m p_4^n}{\vec{b}(m,n)}\right), \] for \(p_1\), \(p_2\), \(p_3\) and \(p_4\) primes and \(\vec b:\mathbb N^2\rightarrow\mathbb N\) satisfying \[ \lim_{m,n\rightarrow\infty}\frac{\vec b(m,n)}{\log(mn)}=0. \] The authors point out that one difference to previously known examples of such sequences is that the components of \(\vec x\) are dependent on both values \(m\) and \(n\) in contrast to known sequences in which the \(j\)-th component was determined by the \(j\)-th component of a generating vector \(\vec v\).