Hardy's inequality for Orlicz-Luxemburg norms (Q810171)

Let \(p>1\), \(x_ n\geq 0\) and \(X_ n=x_ 1+x_ 2+...+x_ n\). Then \[ \sum (X_ n/n)^ p\leq C^ p\sum x^ p_ n \] with \(C=p/(p-1)\). This is the well-known Hardy inequality [cf. Theorem 326 of \textit{G. H. Hardy}, \textit{J. E. Littlewood} and \textit{G. Pólya}: ``Inequalities'' (1974; Zbl 0010.10703); reprint of the 2nd ed. 1988; Zbl 0634.26008)]. This inequality may be written as \[ (1)\quad \| Ax\| \leq C\| x\|, \] where \(x=(x_ n)\) is a sequence treated as a column matrix, \(\| \cdot \|\) is the \(l^ p\)-norm and \(A=(a_{mn})\) is the matrix given by \(a_{mn}=1/m\) if \(0\leq n\leq m,\) \(a_{mn}=0\) otherwise. In the paper the inequality (1) is generalized replacing A by a much more general matrix and \(\|.\|\) by a Luxemburg-type norm with a supermultiplicative Orlicz function.