An analogue to a result of Takahashi (Q321761): Difference between revisions

Let \({\mathcal H}, {\mathcal K}\) and \({\mathcal L}\) denote complex Hilbert spaces. Let \({\mathcal B}(\cdot,\cdot)\) denote the space of bounded linear operators. The following is the main result of the work: Let \(A \in {\mathcal B}({\mathcal H},{\mathcal K})\) and \(C \in {\mathcal B}({\mathcal L},{\mathcal K})\). Then there exists \(X \in {\mathcal B}({\mathcal H},{\mathcal L})\) such that \(A+CX\) is injective if, and only if, \({{\mathcal N}(A)}^{\perp} \cap A^{-1}({\mathcal R}(C))\) contains a closed infinite-dimensional subspace. Here, \({\mathcal N}(\cdot)\) stands for the null space and \({\mathcal R}(\cdot)\) denotes the range space. Applications to a certain completion problem for \(2 \times 2\) operator matrices are studied.