On the closure of the sum of closed subspaces

From MaRDI portal





Taking an arbitrary Hilbert space \({\mathcal H}\) (real or complex) and two closed subspaces \(k,N\subset{\mathcal H}\), the paper under review considers the question whether or not the sum \(K+N=\{k+n; k\in K, n\in N\}\) is closed in \({\mathcal H}\). In a first theorem, three necessary and sufficient conditions for \(K+N\) to be closed are provided. Considering the orthogonal projections \(E_K: {\mathcal H}\to K\), \(E_N: {\mathcal H}\to N\), handy sufficient conditions (e.g., \(E_NE_K=0\)) for \(K+N\) being closed are given in a second theorem. The above results are illustrated by examples, and finally applied to the question whether a solution to an abstract quadratic infinite programming problem exists. Along these lines, sufficient conditions are given such that an optimal solution exists.











This page was built for publication: On the closure of the sum of closed subspaces

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5954784)