A generalization of the Carleson lemma and weighted norm inequalities for the maximal functions in the Orlicz setting (Q2195169)

Let \(\sigma\) be a weight on \(\mathbb R^d\) and \(\alpha\ge 1\). A sequence \(\{\lambda_Q\}_{Q\in\mathcal D}\) of positive numbers indexed over the set of dyadic cubes \(\mathcal D\) in \(\mathbb R^d\) is called a \((\sigma,\alpha)\)-Carleson sequence if there exists a constant \(A>0\) such that or any cube \(R\in\mathcal D\), \[ \sum_{Q\subset R,\ Q\in\mathcal D} \lambda_Q\le A(\sigma(R))^\alpha, \] where \(\sigma(R)=\int_R\sigma(t) dt\). The following equivalent characterization of a \((\sigma,\alpha)\)-Carleson sequence is due to \textit{B. F. Sehba} [J. Math. Soc. Japan 70, No. 1, 71--94 (2018; Zbl 1395.42052)]. Theorem 1.2. Let \(\sigma\) be a weight on \(\mathbb R^d\) and \(\alpha \ge 1\). Assume \(\{\lambda_Q\}_{Q\in\mathcal D} \) is a sequence of positive numbers indexed by \(\mathcal D\). Then the following are equivalent: \par (i) \(\{\lambda_Q\}_{Q\in \mathcal D}\) is a \((\sigma,\alpha)\)-Carleson sequence. \par (ii) There exists a constant \(B>0\) such that for all \(p\in(1,\infty)\), \[ \sum_{Q\in\mathcal D} \lambda_Q[m_\sigma(f,Q)]^{p \alpha} \le B\|f\|_{p,\sigma}^{p\alpha}, \] where \(m_\sigma(f,Q)= \frac1{\sigma(Q)}\int_Q f(t)\sigma(t)dt\). \par When \(\alpha =1\) Theorem 1.2 is the classical Carleson's lemma. \par The first aim of the paper is to provide an extension of Theorem 1.2 to the case when the functions \(t\to t^{p\alpha},\, \alpha \ge 1\), are replaced by Youngs functions of an appropriate subclass. In Definition 2.1 the authors define a \((\sigma,\Phi)\)-Carleson sequence where \(\sigma\) is a weight on \(\mathbb R^d\) and \(\Phi\) is a growth function on \([0,\infty)\). In Theorem 2.2 they provide a generalization of Theorem 2.1 to the case where \(\Phi_1\) and \(\Phi_2\) are growth functions satisfying certain conditions. This result is then applied to obtain two-weight norm estimates for some maximal functions between different Orlicz spaces.