Continuous homogeneous selections of set-valued metric generalized inverses of linear operators in Banach spaces (Q1943012)

Let \(X\) and \(Y\) be real Banach spaces and \(T: X \to Y\) be a linear operator. An element \(x_0 \in X\) is called an extremal solution of the equation \(Tx = y\) if \(x = x_0\) minimizes the value of \(\|Tx - y\|\). An extremal solution with minimal norm is called a best approximation solution. According to \textit{M. Z. Nashed} and \textit{G. F. Votruba} [Bull. Am. Math. Soc. 80, 825--830, 831--835 (1974; Zbl 0289.47011)], if \(y \in Y\) such that \(Tx=y\) has a best approximation in \(X\), then \(T^\partial(y)\) is defined as the set of all \(x \in X\) such that \(x\) is a best approximation solution to \(Tx=y\), and the set-valued mapping \(y \mapsto T^\partial(y)\) is called the metric generalized inverse. For the selection of the metric generalized inverse of linear operator in Banach space, \textit{H. Hudzik}, \textit{Y.-W. Wang} and \textit{W.-J. Zheng} [Set-Valued Anal. 16, No. 1, 51--65 (2008; Zbl 1148.47003)] obtained a bounded homogeneous selection. In the paper under review, the authors investigate continuous homogeneous selections for \(T^\partial\) by means of the methods of geometry of Banach spaces. They present some necessary and sufficient conditions for bounded linear operators \(T\) to have continuous homogeneous selections for \(T^\partial\).