Left invertible completions of upper triangular operator matrices with unbounded entries (Q2355351): Difference between revisions

Let \(X\) and \(Y\) be separable Hilbert spaces, \({\mathcal B}(X,Y)\) denote the space of all bounded linear operators from \(X\) into \(Y\), and \({\mathcal C}(X,Y)\) denote the space of all closed linear operators \(X\) into \(Y\). If \(X=Y\), then \({\mathcal B}(X,Y)\) is abbreviated to \({\mathcal B}(X)\) and \({\mathcal C}(X,Y)\) is abbreviated to \({\mathcal C}(X)\). For a linear operator \(T\) from \(X\) into \(Y\), the domain, the range space and the null space are denoted by \({\mathcal D}(T),~{\mathcal R}(T)\) and \({\mathcal N}(T)\), respectively. Let \(n(T)\) denote the nullity of \(T\) and \(d(T)\) stand for the dimension of the orthogonal complement of \({\mathcal R}(T)\). The main result of the authors is the following: Let \(A \in {\mathcal C}(X)\) with \(d(A)\) finite and \(C \in {\mathcal C}(Y)\). Then the upper triangular operator matrix \(\left(\begin{matrix} A & B \\ 0 & C \end{matrix} \right):{\mathcal D}(A) \oplus {\mathcal D}(C) \rightarrow X \oplus Y\) is a left invertible closed operator for some closable operator \(B\) from \(Y\) into \(X\) if and only if \(A\) is left invertible with \(n(C) \leq d(A)\) and \(R(C)\) is closed. In the case when the orthogonal complement of \({\mathcal R}(T)\) is infinite dimensional, it is shown that the upper triangular operator matrix given as above is left invertible if and only if so is \(A\).