Linear preservers of weak majorization on \(\ell^{1}(I)^{+}\), when \(I\) is an infinite set (Q503426): Difference between revisions

Let \(I\) and \(J\) be arbitrary non-empty sets and let \(p \in \left [1 ,\infty \right )\). Denote by \(l^{p}(I)\) the discrete Lebesgue space and by \(e_{i} :I \rightarrow \mathbb{R}\) functions defined by the Kronecker delta, i.e., \(e_{i}(j) =\delta _{ij}\), \(i ,j \in I\). The weak majorization relation is defined on the set of all positive elements of the Banach space \(l^{p}(I)\), i.e., \(l^{p}(I)^{ +} =\{f \in l^{p}(I) :f(i) \geq 0 \), for every \(i \in I\}\). A bounded linear operator \(A :l^{p}(J) \rightarrow l^{p}(I)\) is called {\parindent=0.8cm \begin{itemize}\item[{\(\bullet\)}] positive, if \(Ag \in l^{p}(I)^{ +}\) for each \(g \in l^{p}(J)^{ +}\); \item[{\(\bullet\)}] doubly substochastic, if \(A\) is positive, \(\sum _{j \in J}\left \langle Ae_{j} ,e_{i}\right \rangle \leq 1\) for every \(i \in I\), and \(\sum _{i \in I}\left \langle Ae_{j} ,e_{i}\right \rangle \leq 1\) for every \(j \in J\). \end{itemize}} Let \(f ,g \in l^{p}(I)^{ +}\). We say that \(f\) is weakly majorized by \(g\) if there exists a doubly substochastic operator \(D\) on \(l^{p}(I)\) such that \(f =Dg\), and write \(f \prec _{w}g\). A bounded linear operator \(T :l^{p}(I) \rightarrow l^{p}(I)\) is called a preserver of weak majorization on \(l^{p}(I)^{ +}\) if \(f \prec _{w}g\), where \(f ,g \in l^{p}(I)^{ +}\), implies \(Tf \prec _{w}Tg\). The set of all linear preservers of weak majorization on \(l^{p}(I)^{ +}\) is denoted by \(\mathcal{M}_{wpr}(l^{p}(I)^{ +})\). In the paper, the authors mostly focus on the case \(p =1\). Namely, they study linear preservers of weak majorization on \(l^{1}(I)^{ +}\). Let from now on \(I\) be an infinite set. The authors present necessary and sufficient conditions for a bounded linear map on \(l^{1}(I)\) to be a preserver of weak majorization on \(l^{1}(I)^{ +}\). For example, they prove that \(T \in \mathcal{M}_{wpr}(l^{1}(I)^{ +})\) if and only if \(Te_{j} \prec _{w}Te_{k}\) and \(Te_{k} \prec _{w}Te_{j}\), for every \(k ,j \in I\), and for each \(i \in I\), there either exists exactly one \(j \in I\) with \(\left \langle Te_{j} ,e_{i}\right \rangle >0\) or the set \(\left \{\left \langle Te_{j} ,e_{i}\right \rangle :j \in I\right \}\) is a singleton. The authors conclude the paper by showing that the set \(\mathcal{M}_{wpr}(l^{1}(I)^{ +})\) is a norm-closed subset of the set of all bounded linear operators on \(l^{1}(I)\).