Transitive spaces of operators (Q939145)

A collection \(\Sigma\) of operators from a vector space \(X\) to another vector space \(Y\) is said to be transitive if, given nonzero \(x \in X\) and \(y \in Y\), there is an element \(A \in \Sigma\) such that \(Ax=y\). This paper represents an effort to extract as much information as possible from transitivity from the linear structure on \(\Sigma\) alone. Starting with finite dimensions, the authors present a number of positive results and a lot of counterexamples to natural conjectures. Among other things, they determine the minimal dimension of \(k\)-transitive subspaces of \(B(H,K)\) in terms of the dimensions of \(H\) and \(K\). They consider spans of products as well as tensor products of spaces, and study the relations between their degree of transitivity and those of their constituent spaces. They investigate the relations between the minimal and maximal ranks present in a transitive subspaces of \(M_{n}\). In particular, they show that such a subspace must contain invertible elements. Also, they extend some of the finite-dimensional results.