On pair correlation and discrepancy (Q2407399): Difference between revisions

The authors of this paper study infinite sequences \((x_n)_{n\geq 1}\) in \([0,1)\). Such a sequence is said to have Poissonian pair correlations if \[ \lim_{N\rightarrow\infty}\frac{1}{N} \# \left\{1\leq l\neq m\leq N:\, \| x_l - x_m \|<\frac{s}{N}\right\}=2s \] for all positive \(s\). Furthermore, a sequence \((x_n)_{n\geq 1}\) in \([0,1)\) is called uniformly distributed if \[ \lim_{N\rightarrow\infty}\frac{1}{N} \# \left\{1\leq n\leq N:\, x_n\in [a,b)\right\}=b-a \] for all \(0\leq a\leq b\leq 1\). The main result in the paper links the property of having Poissonian pair correlations to the property of uniform distribution in the following sense. It is known that it is not necessary for a sequence to have Poissonian pair correlations in order to be uniformly distributed. However, the main result in the present paper shows that the converse is true, i.e., if a sequence has Poissonian pair correlations then it is also uniformly distributed. This result is shown by linking Poissonian pair correlations to the so-called star discrepancy, a measure of uniform distribution.