The representation and approximations of outer generalized inverses (Q558248)

Let \(X\) and \(Y\) be Banach spaces, and \(A\in L(X,Y)\). An operator \(G\in L(X,Y)\) is called an outer generalized inverse \((OGI)\) of \(A\) if \(GAG=G\). A unified representation theorem for the class of all \(OGI\)'s of an operator is presented. The theorem is a generalization for the corresponding representation of the Moore-Penrose inverse [see \textit{C. W. Groetsch}, J. Math. Anal. Appl. 49, 154-157(1975; Zbl 0295.47012)], of the Drazin inverse [see \textit{Y. Wei} and \textit{S. Qiao}, Appl. Math. Comput. 138, 77-89 (2003; Zbl 1034.65037)], and of the specific generalized inverse studied by \textit{Y. Wei} [Linear Algebra Appl. 280, 87-96 (1998; Zbl 0934.15003)]. The unified representation is used to develop several particular expressions and computational procedures for the set of \(OGI\)'s. Some illustrative numerical examples are given.