On disjoint range operators in a Hilbert space (Q448387): Difference between revisions

Let \(\mathcal H\) and \(\mathcal K\) be complex Hilbert spaces and \(\mathcal B(\mathcal H, \mathcal K)\) denote the set of all bounded linear operators from \(\mathcal H\) into \(\mathcal K\). \(\mathcal B(\mathcal H)\) stands for \(\mathcal B(\mathcal H, \mathcal K)\) when \(\mathcal H = \mathcal K\). For \(M \in \mathcal B(\mathcal H, \mathcal K)\), \(M^*, {\mathcal R} (M)\) and \({\mathcal N} (M)\) will denote the adjoint, the range and the null space of \(M\), respectively. In the paper under review, the authors consider various relations between the range spaces of \(M, M^*, M+M^*\) and the null spaces of \(M\) and \(M^*\), in the context of their relationships with several special classes of operators, namely, EP, co-EP, weak-EP, etc. Let us give the definition for one of these classes: Let \(M \in \mathcal B(\mathcal H, \mathcal K)\) be such that \({\mathcal R} (M)\) is closed. Then \(M\) is said to be an EP-operator if \({\mathcal R}(M)={\mathcal R}(M^*)\). In what follows, we present a sample result. \(MM^{\dagger}-M^{\dagger}M\) is EP if and only if \(MM^{\dagger}(I-M^{\dagger}M)M^{\dagger}M\) is EP if and only if \(MM^{\dagger}(I-M^{\dagger}M)\) is Moore-Penrose invertible. Here, an operator \(B \in \mathcal B (\mathcal H, \mathcal K)\) is said to be Moore-Penrose invertible if there exists \(C \in \mathcal B (\mathcal K, \mathcal H)\) such that \(BCB=B\), \(CBC=C\), \((BC)^*=BC\) and \((CB)^*=CB\). When \(M\) is represented as a block operator, equivalent statements on the block entries are also presented. In the process, several known results for matrices are extended to operators between Hilbert spaces.