English

Generalized Discrepancy of Random Points

Numerical Analysis 2025-12-10 v1 Numerical Analysis Number Theory Probability

Abstract

We study the LpL_p-discrepancy of random point sets in high dimensions, with emphasis on small values of pp. Although the classical LpL_p-discrepancy suffers from the curse of dimensionality for all p(1,)p \in (1,\infty), the gap between known upper and lower bounds remains substantial, in particular for small p1p \ge 1. To clarify this picture, we review the existing results for i.i.d.\ uniformly distributed points and derive new upper bounds for \emph{generalized} LpL_p-discrepancies, obtained by allowing non-uniform sampling densities and corresponding non-negative quadrature weights. Using the probabilistic method, we show that random points drawn from optimally chosen product densities lead to significantly improved upper bounds. For p=2p=2 these bounds are explicit and optimal; for general p[1,)p \in [1,\infty) we obtain sharp asymptotic estimates. The improvement can be interpreted as a form of importance sampling for the underlying Sobolev space Fd,qF_{d,q}. Our results also reveal that, even with optimal densities, the curse of dimensionality persists for random points when p1p\ge 1, and it becomes most pronounced for small pp. This suggests that the curse should also hold for the classical L1L_1-discrepancy for deterministic point sets.

Keywords

Cite

@article{arxiv.2512.08364,
  title  = {Generalized Discrepancy of Random Points},
  author = {Erich Novak and Friedrich Pillichshammer},
  journal= {arXiv preprint arXiv:2512.08364},
  year   = {2025}
}
R2 v1 2026-07-01T08:16:26.788Z