Bounds for \(L_p\)-discrepancies of point distributions in compact metric measure spaces (Q2309576)

Uniformly distributed point sets have many applications to different branches of the science, especially to the discrete geometry. These applications have been studied by many authors. In the present paper, for all exponents \(0 < p < \infty\) and \(p=\infty\) the author presents upper bounds for the \(L_p\)-discrepancy of point distributions in compact metric measure spaces, under very simple conditions on the volume of the metric balls. The obtained upper bounds are of the sharp order, at least, for \(2 \leq p < \infty\) and Riemannian symmetric manifold of rank one. In the introduction of the paper, some notations and confirmations of the distribution theory in compact metric measure spaces are reviewed. Here \(\mathcal{M}\) is a compact metric measure space with a fixed metric \(\theta\) and a finite non-negative Borel measure \(\mu\). The concept of the local discrepancy of a \(N\)-point subset \(\mathcal{D}_N \subset \mathcal{M}\) is reminded. For \(0 < p < \infty\) and \(p=\infty\) the \(L_p\)-discrepancy of the set \(\mathcal{D}_N\) are defined. Two kinds of the extremal discrepancy \(\lambda_p[\xi, N]\) and \(\lambda_{\infty}[N]\) concerning geometric characteristics of the space \(\mathcal{M}\) are introduced. Also, two-side bounds of \(\lambda_2[\xi, N]\) are presented. Some upper bounds of the \(L_p\)-discrepancy and \(L_{\infty}\)-discrepancy are reminded and discussed. In Section 2, the main results of the paper are presented. First, metric measure spaces are specialized by two conditions. The first condition is related with Ahlfors regular spaces and the second one considers the Lipschitz continuity of the volume \(v(x, y)\). In Theorem 2.1 \(\mathcal{M}\) is a compact connected metric measure space satisfying the introduced conditions. For \(0 < p < \infty\) the upper bound of the extremal discrepancy \[ \lambda_p[\mathcal{M}, \xi, N] \leq c(p+1)^{\frac{1}{2}} N^{\frac{1}{2} - \frac{1}{2d}}, \] where \(\xi\) is an arbitrary normalized measure for any compact Riemannian manifold of dimension \(d\), is obtained. In Proposition 2.2 under the hypothesis of Theorem 2.1, the author finds an upper bound of the \(L_{\infty}\)-discrepancy \[ \mathcal{L}_{\infty}[\mathcal{M}, \mathcal{D}_N] \leq 2 m^{\frac{2}{p}} \mathcal{L}_{p}[\mathcal{M}, \xi_0, \mathcal{D}_N] + c_4 N m^{-\frac{1}{d}}, \] where \(\xi_0\) is the standard Lebesgue measure, of an arbitrary \(N\)-point subset of \(\mathcal{M}\). In Corollary 2.1 under the hypothesis of Theorem 2.1 for all \(N\), the author obtains an upper bound of the extremal discrepancy \[ \lambda_{\infty}[\mathcal{M}, N] \leq c N^{\frac{1}{2} - \frac{1}{2d}}(\log N)^{\frac{1}{2}}. \] In Section 3, necessary preliminary facts on partitions of metric measure spaces are developed. Also, Proposition 2.2 is proved. In Section 4, the construction of random point distributions is described and Theorem 2.1 is proved. In Section 5 Proposition 2.1 is proved.