Central limit theorem for \((t,s)\)-sequences in base 2 (Q2685063)

Let \((x_n)\) be a digital \((t,s)\)-sequence in base 2, and let \(\mathbb{P}_m\) denote the set of the first \(2^{m}-1\) elements of \((x_n)\). Let \(Y\) denote a random variable giving a random point \((y_1, \ldots, y_s) \in [0,1]^s\) and let \(T\) denote a random shift with \(T \in \mathbb{Q}^s(2^m)\) for a given \(m\). The author derives in Theorem 1 a central limit theorem for digital \((t,s)\)-sequences in base 2 of the following form. \[ 2^{-sm} \sum_{T \in \mathbb{Q}^s(2^m)} \mathrm{vol}\left \{ Y \in [0,1)^s : \frac{D(\mathbb{P}_m \oplus T, Y)}{\mathcal{M}_{s,2}(\mathbb{P}_m)} < w \right\} \overset{m \rightarrow \infty} \longrightarrow \Phi(w). \] Here \(D(.,.)\) denotes the local discrepancy function, while \[ \mathcal{M}_{s,p}(\mathbb{P}_m)=\left( \int_{[0,1)^{2s}} |D(\mathbb{P}_m \oplus T,Y)|^p dT dY\right)^{1/p}, \] in which \(D_p(.)\) is the \(L_p\)-discrepancy and \(0<p<\infty\). Moreover, the author proves in Theorem 2 that \[ \frac{\mathcal{M}_{s,p}(\mathbb{P}_m)}{\mathcal{M}_{s,2}(\mathbb{P}_m)} \overset{m \rightarrow \infty} \longrightarrow \chi_p, \] in which \[ \chi_p=\frac{1}{\sqrt{2\pi}} \int_{\infty}^{\infty} |u|^p \exp(-u^2/2) du. \]