Extension theorems for vector valued maps (Q2490003): Difference between revisions

The authors consider two extension problems motivated by some questions in optimal design. The first problem is about extensions of Lipschitz maps between Banach spaces. Let \(E\) and \(F\) be Banach spaces and \({D\subset E}\). Does a map \({u:D\to F}\) satisfying \[ \| u(x)-u(y)\| _F\leq \| x-y\| _E,\;\;x,y\in D, \] admit an extension to the whole of \(E\) so as to preserve the inequality? This problem was considered by Kirszbraun, McShane, Valentine, Grünbaum, Minty, and Schön\-beck. Their research was summarized in the book [\textit{J. H. Wells} and \textit{L. R. Williams}, ``Embeddings and extensions in analysis'' (Ergebnisse der Mathematik und ihrer Grenzgebiete 84, Springer-Verlag) (1975; Zbl 0324.46034)]. The main result is that if \({\dim E, \dim F \geq 2}\) and the unit sphere of \(F\) is strictly convex, then the pair \([E;F]\) has the extension property for contractions if and only if both \(E\) and \(F\) are Hilbert spaces. In the paper under review, the authors give a different, more elementary and more self-contained proof of this result. The main tool used by the authors is the smallest norm above \(\| \cdot\| _E\) which is induced by an inner product. In the special case when the initial map \(u\) is defined on a closed convex set \(\Omega\), the following is proved. Assume that \(E\) is a Hilbert space and \(F\) is a normed space. Then every contraction \({u:\Omega\to F}\) has an extension \({\tilde u:E\to F}\) that is still a contraction. The second problem considered is the following. Let \({D\subset {\mathbb R}^d}\) and suppose that \({u:D\to {\mathbb R}^d}\) satisfies \[ | \langle u(x)-u(y);x-y \rangle | \leq \| x-y\| ^2,\;\;x,y\in D. \] Does the map \(u\) admit an extension to the whole of \({\mathbb R}^d\) so as to preserve the last inequality? If the answer is positive for every \(u\), we say that \(D\) has the extension property for displacements of bounded strains. The family of subsets with this property is very poor. For instance, the authors prove that any convex set of nonempty interior in \({\mathbb R}^2\) does not have the extension property for displacements of bounded strains.