Sequences with almost Poissonian pair correlations (Q2116745)

An infinite sequence \((x_n)\) in \([0,1]\) has \(\alpha\)-pair correlations for \(0<\alpha\leq 1\) if \[\lim_{N \rightarrow \infty} \frac{1}{N^{2-\alpha}} \# \left \{ 1 \leq l \neq m \leq N : \|x_l - x_m \| \leq \frac{s}{N^{\alpha}} \right \} =2s.\] In the special case \(\alpha=1\) this coincides with the well-known notion of Poissonian pair correlations. An i.i.d uniform random sequence has almost surely Poissonian pair correlations and having Poissonian pair correlations implies that a sequence is uniformly distributed [\textit{C. Aistleitner} et al., J. Number Theory 182, 206--220 (2018; Zbl 1415.11106); \textit{S. Grepstad} and \textit{G. Larcher}, Arch. Math. 109, No. 2, 143--149 (2017; Zbl 1387.11064)]. However, it is known that many classical examples of uniformly distributed sequences such as the van der Corput sequence or the Kronecker sequence do not have Poissonian pair correlations. The main results of this paper are the observations that the van der Corput sequence as well as the Kronecker sequence generated from the golden mean are as close to having Poissonian pair correlations as possible, i.e., both sequences have \(\alpha\)-pair correlations for all \(0<\alpha<1\).