English

Optimal Polynomial Tractability Exponents for the Inverse Star Discrepancy

Numerical Analysis 2026-07-26 v1

Abstract

The inverse of the star discrepancy n(d,ε)n^\ast(d, \varepsilon) satisfies dε1n(d,ε)dε2 d \varepsilon^{-1} \lesssim n^{\ast}(d,\varepsilon)\lesssim d\varepsilon^{-2} for all dNd \in \mathbb{N} and 0<ε<ε00 < \varepsilon < \varepsilon_0. The upper bound was shown by Heinrich, Novak, Wasilkowski and Wo\'{z}niakowski (2001) and the lower bound was shown by Hinrichs (2004). A proof of this lower bound using elementary arguments was shown by Steinerberger (2023). These bounds show that the inverse of the star discrepancy depends linearly on the dimension, but the exact exponent of ε1\varepsilon^{-1} has remained open. In this paper we prove a lower bound which shows that the exponent 22 of ε1\varepsilon^{-1} in the upper bound cannot be improved. More precisely, we show that for every 0<α<10<\alpha<1 and fixed 0<AB0<A\le B, there are constants cα,B>0c_{\alpha,B}>0 and εα,A>0\varepsilon_{\alpha,A}>0 such that for every integer dd satisfying AεαdBεα A\varepsilon^{-\alpha}\le d\le B\varepsilon^{-\alpha} we have n(d,ε)cα,Bdε(2α), n^{\ast}(d,\varepsilon) \ge c_{\alpha,B}\,d\,\varepsilon^{-(2-\alpha)}, for all 0<ε<εα,A0 < \varepsilon < \varepsilon_{\alpha, A}. Along these polynomial strips the right-hand side is of order ε2\varepsilon^{-2}. It follows that every uniform polynomial upper estimate n(d,ε)Cdqεpn^{\ast}(d,\varepsilon)\le C d^q\varepsilon^{-p} must satisfy p2p \ge 2. Together with the lower bound of Hinrichs (2004), which forces q1q \ge 1, this shows that the exponents p=2p=2 and q=1q=1 in the Heinrich--Novak--Wasilkowski--Wo\'{z}niakowski upper bound are individually optimal.

Cite

@article{arxiv.2607.23571,
  title  = {Optimal Polynomial Tractability Exponents for the Inverse Star Discrepancy},
  author = {Josef Dick},
  journal= {arXiv preprint arXiv:2607.23571},
  year   = {2026}
}

Comments

10 pages, 1 figure