The maximal quartile operator (Q5940350): Difference between revisions

Let \(C^\Delta(\mathbb{R})\) denote the set of all functions \(f: \mathbb{R}\to\mathbb{R}\) that are finite linear combinations of characteristic functions of dyadic intervals, i.e., intervals of the form \([2^k n,2^k(n+ 1))\) with \(k,n\in\mathbb{Z}\). We define the \textit{Walsh function} \(W_l\in C^\Delta(\mathbb{R})\) for \(l\in\mathbb{N}_0\) by the following recursive formulas \[ W_0= \mathbf{1}_{[0,1)}, \] \[ W_{2l}= W_l(2x)+ W_l(2x- 1), \] \[ W_{2l+ 1}= W_l(2x)- W_l(2x- 1). \] For \(k,n\in \mathbb{Z}\), \(l\in \mathbb{N}_0\) we define the \textit{Walsh wave packet} \(w_{k,n,l}\) by \[ w_{k,n,l}(x)= 2^{-k/2} W_l(2^{-k}x- n). \] The quartile operator \(H_W\) and the maximal quartile operator \(H^{\max}_W\) are then defined by \[ H_W(f, g):= \sum_{\substack{ k,n\in\mathbb{Z}\\ l\in\mathbb{N}_0}} 2^{-k/2}\langle f,w_{k,n,4l}\rangle \langle g, w_{k,n,4l+ 1}\rangle w_{k,n,4l+ 2}, \] \[ H^{\max}_W(f, g)(x):= \sup_{K\in \mathbb{Z}} \Biggl|\sum_{\substack{ k,n\in \mathbb{Z}\\ l\in\mathbb{N}_0\\ k\leq K}} 2^{-k/2}\langle f, w_{k,n,4l}\rangle \langle g,w_{k,n,4l+ 1}\rangle w_{k,n,4l+ 2}(x)\Biggr|. \] In this paper, the author proves the following theorem. Theorem 1. Let \(p\), \(q\), \(r\) satisfy \({1\over p}={1\over q}+{1\over r}\), \({2\over 3}< p< \infty\), \(1< q,r\leq \infty\). Then there is a constant \(C\) such that for all functions \(f,g\in C^\Delta(\mathbb{R})\) \[ \|H_W(f, g)\|_p\leq C\|f\|_q\|g\|_r, \] \[ \|H^{\max}_W(f, g)\|_p\leq C\|f\|_q\|g\|_r. \]