Linear surjective isometries between vector-valued function spaces

From MaRDI portal





Let \(X,Y\) be compact Hausdorff spaces with \(X\) homotopically rigid and \(E\) a smooth, reflexive Banach space. In this paper, the author investigates the structure of surjective linear isometries of subspaces of \(A \subset C(X,E)\) and \(B \subset C(Y,E)\), under some additional assumptions on the subspaces. Assume that, for each \(u \in E\), the constants maps are in the subspace and for distinct points \(x_1,x_2 \in X\), there is a function \(f \in A\) with \(f(x_1) = u\) and \(f(x_2) = 0\) (also, a similar assumption involving \(Y\) and \(B\)). Assume further that both \(A,B\) satisfy the condition that, if \(x \in X\) (or \(y \in Y\)) is such that \(e^\ast \circ \delta(x)\) (or \(e^\ast \circ \delta(y)\)) is an extreme point of the dual unit ball, for some extreme point \(e^*\) of the dual unit ball, then the same conclusion holds for all extreme points of the dual unit ball (such a set of points is called the Choquet boundary of the subspace). Then, Theorem 3.5, as in the classical Banach-Stone theorems [\textit{E. Behrends}, M-structure and the Banach-Stone theorem. Berlin-Heidelberg-New York: Springer-Verlag (1979; Zbl 0436.46013)], any surjective linear isometry \(T: A \rightarrow B\) is a composition operator involving a homeomorphism from \(Y\) to \(X\) and a continuous family of linear isometries on \(E\) indexed by the Choquet boundary of \(B\).











This page was built for publication: Linear surjective isometries between vector-valued function spaces

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2811984)