Sequences of dilations and translations equivalent to the Haar system in \(L^p\)-spaces (Q2071755)

Set for convenience \(L^p = L^p[0,1]\) and \(L^p_0 := \{x \in L^p: \, \int_{[0,1]} x \, d \mu = 0\}\), \(1 \le p < \infty\). By \((x_n) \stackrel{p}{\sim} (y_n)\) we mean that the corresponding sequences are equivalent in \(L^p\). The \emph{sequence of dilations and translations} (s.d.t. in short) of a function \(f \in L^1_0\) is defined by setting \(f_n(t) = f(2^k t - j)\) for \(t \in I_k^j := \left(\frac{j}{2^k}, \frac{j+1}{2^k} \right]\) and \(f_n(t)=0\) for all \(t \in [0,1] \setminus I_k^j\), where \(n = 2^k+j\), \(k = 0,1, \ldots\) and \(j = 0, \ldots, 2^k-1\). So the usual (\(L_\infty\)-normalized) Haar system \((h_n)_{n=1}^\infty\) without its starting element \(h_0 = \chi_{[0,1]}\) coincides with the s.d.t. of \(h := \chi_{I_1^0} - \chi_{I_1^1}\). For every \(d \in \mathbb N\), let \(N_d\) be the set of all \(n \in \mathbb N\) of the form \(n = 2^{i_1}+\ldots+2^{i_d}\) for suitable integers \(i_1 > \ldots > i_d \ge 0\). Every element \(x \in L^1_0\) of the form \(x = \sum_{n \in N_d} \xi_n h_n\), where \(\xi_n \in \mathbb C\), is called a \emph{Haar chaos of order} \(d \in \mathbb N\), and the set of all such elements is denoted by \(\mathcal{H}_{ch}^d\). The paper under review is mostly devoted to the characterization of elements \(f \in L^p_0\) such that the s.d.t. \((f_n)_{n=1}^\infty\) of \(f\) satisfies \((f_n)_{n=1}^\infty \stackrel{p}{\sim} (h_n)_{n=1}^\infty\). The most interesting results (which are essentially different depending on whether \(p \ge 2\) or \(p < 2\)) are obtained for the case where \(f\) is a Haar chaos of order~\(1\). For instance, Theorem~1 asserts the equivalence of the following conditions on a Haar chaos $f$ of order~\(1\) for \(p \ge 2\): (i) \([f_n] = L^p_0\) and \((f_n)_{n=1}^\infty \stackrel{p}{\sim} (h_n)_{n=1}^\infty\). (ii) For each \(d \in \mathbb N\) one has \([f_n]_{n \in N_d} = \mathcal{H}_{ch}^d \cap L^p\) and \((f_n)_{n \in N_d} \stackrel{p}{\sim} (h_n)_{n \in N_d}\). (iii) The same two conditions as in (ii), but for \(d=1\) only. Moreover, each of the conditions (i), (ii), (iii) is equivalently characterized by the authors in terms of analytic behavior of the so-called generating function \(\hat{f}\) of \(f\).