Variants of the Maurey-Rosenthal theorem for quasi Köthe function spaces (Q5936413)

According the Maurey-Rosenthal theorem each bounded linear operator \(T: E\to L_p(\nu)\), \(0<p<r<\infty\) from a quasi normed space \(E\), which allows the estimate \[ \left\|\left(\sum|Tx_k|^r\right)^\frac 1r\right\|_{L_p}\leq \left(\sum\|Tx_k\|^r_E\right)^\frac 1r\quad \text{ for all}\quad x_1,\ldots,x_n\in E , \] allows a similar estimate \[ \left(\int\frac{|Tx|^r}{w}d\nu\right)^\frac 1r\leq \|x\|_E \quad {\text{ for all}}\quad x\in E \] with some weight function \(0\leq w\in L_0(\nu)\). The author extends this result to the case of homogeneous operators. The following result is the main one in the paper. Theorem 2. Let \(T:U\to V\) be a homogeneous operator where \(U\) and \(V\) are vector spaces (or only homogeneous sets) which via \(\phi\) and \(\psi\) are represented homogeneously in quasi Köthe function spaces \(X(\mu)\) and \(Y(\nu)\), respectively. For \(0<r<\infty\) let \(X(\mu)\) be \(r\)-convex, and \(Y(\nu)\) \(r\)-concave. If \(T\) satisfies \[ \left\|\left(\sum\left|\psi(Tx_k)\right|^r\right)^\frac 1r\right\|_Y\leq \left\|\left(\sum|\phi(x_k)|^r\right)^\frac 1r\right\|_X \] for all \(x_1,\ldots,x_n\in U\), then there are \[ \begin{aligned} 0\leq\varphi: X^r\rightarrow \mathbb{R} \quad{ \text{linear with}}\quad&\sup_{\|x\|_X\leq 1} \varphi(|x|^r)^\frac 1r\leq M^{(r)}(X)\\ 0\leq\omega_2\in L_0(\nu)\quad {\text{with}}\quad &\sup_{\|y\|_{L_r(\nu)}\leq 1}\|\omega^\frac 1r_2 y\|_Y\leq M_{(r)}(Y) \end{aligned} \] such that for all \(x\in U\) \[ \int\frac{|\psi(Tx)|^r}{\omega_2}d\nu\leq \varphi(|\phi(x)|^r) . \] If \(X(\mu)\) is \(\sigma\)-order continuous, then there is even a function \[ 0\leq\omega_1\in L_0(\mu)\quad \text{with}\quad \sup\limits_{\|x\|_X\leq 1}\|\omega^\frac 1r_1 x\|_{L_r(\mu)}\leq M_{(r)}(X) \] such that for all \(x\in U\) \[ \int\frac{|\psi(Tx)|^r}{\omega_2}d\nu\leq\int|\phi(x)|^r \omega_1d^\mu . \] Here \(M_{(r)}(X)\) and \(M^{(r)}(X)\) stand for the best constants in the inequalities \[ \left(\sum\|x_k\|^r_X\right)^\frac 1r\leq c \left\|\left(\sum|x_k|^r\right)^\frac 1r\right\|_X \qquad \text{ and} \qquad \left\|\left(\sum|x_k|^r\right)^\frac 1r\right\|_X\leq c \left(\sum\|x_k\|^r_X\right)^\frac 1r. \]