Norm estimates for operators from \(H^p\) to \(\ell^q\) (Q924109): Difference between revisions

Let \(1\leq p,q\leq\infty\) and \(T:H^{p}\rightarrow l^{q}\) be a continuous linear operator, where \(H^{p}\) is the Hardy space in the unit disc. For every integer \(n\geq0\), let \(u_{n}\in H^{p}\) be given by \(u_{n}(z)=z^{n}\) and let \((e_{k})_{k=1}^{\infty}\) denote the canonical vectors of \(l^{q}\) (canonical Schauder basis when \(1\leq q<\infty\)). Consider the matrix \((t_{kn})_{k,n}\) defined by \[ T(u_{n})= {\sum_{k=1}^{\infty}}t_{kn}e_{k} \] and let \(T_{k}=(t_{kn})_{n\geq0}\). \textit{B.\,Osikiewicz} [Quaest.\ Math.\ 28, No.\,3, 263--270 (2005; Zbl 1097.46014)] obtained several theorems concerning upper and lower estimates for the norm \(\left\| T\right\|\) in terms of \(\left\|(T_{k})\right\|_{l^{r}(l^{s})}\) for different values of \(r\) and \(s\). In the paper under review, the author improves the results from [loc.\,cit.] by using not only the norms of the rows \(T_{k}\) but also the norms of the columns \(x_{n} =(t_{kn})_{k\geq1}\). The author also considers \(l(p,q)\)-spaces instead of \(l^{q}\)-spaces to obtain more precise estimates. The main tool used in the paper is the description of the boundedness of operators from \(H^{p}\) to \(l^{q}\) by means of vector-valued functions.