On mean values of the \(L_q\)-discrepancies of point distributions (Q2856443)

Let \(D_N\) be an arbitrary set of \(N>1\) points in the \(d\)-dimensional unit cube \([0,1)^d\) and let NEWLINE\[NEWLINE \mathcal{L}[D_N, Y] = \#\{ D_N \cap [0,Y) \} - N \mathrm{vol}[0,Y) NEWLINE\]NEWLINE be the local discrepancy, in which \([0,Y)\) is a subinterval of \([0,1)^d\) anchored at 0. Then the \(L_q\)-discrepancy is defined as the \(L_q\)-norm of the local discrepancy for \(1 \leq q < \infty\).NEWLINENEWLINEThe paper is a continuation of an earlier paper of the same author [St. Petersbg. Math. J. 23, No. 4, 761--778 (2012); translation from Algebra Anal. 23, No. 4, 179--204 (2011; Zbl 1278.11077)] and considers digital \((\delta, s, d)\)-nets in base 2. Staying within the technical framework of the first paper, the author introduces the concept of \textit{conditional mean \(L_q\)-discrepancies} and extends his upper bound results obtained for the mean values of the \(L_q\) discrepancies to this more general setting.NEWLINENEWLINEAs a reward for the additional effort, the author obtains many interesting corollaries presented in Section 3 and 4 showing that various results on uniform distribution follow directly from his deep results on the mean values of the \(L_q\)-discrepancies.