\(C^1\) extensions of functions and stabilization of Glaeser refinements (Q2468416): Difference between revisions

In 1934 Whitney gave a complete description of traces of \(C^m\)-functions on an arbitrary compact set \(E\) in the real line \(\mathbb{R}\). A procedure for determining whether a \(C^1\) extension of a function \(f:E \rightarrow \mathbb{R}, E \subset \mathbb{R}^n,\) exists was obtained by Glaeser in 1958. He introduced the concept of paratangent bundle. \textit{C. Fefferman} gave in a series of papers since 2003 (in particular [Ann. Math. 164, 313--359 (2006; Zbl 1109.58016)]) a remarkable extension of Whitney's theorem to higher dimensions and higher degrees of smoothness for an arbitrary compact subset of \(\mathbb{R}^n\). A key ingredient of Fefferman's description of traces of \(C^m\) functions on compact subsets is the notion of Glaeser refinement; a concept related to the iterated paratangent bundles. The authors of this paper study Glaeser refinements in the case of smoothness one and their stabilization properties. Rather complete answers are given for \(C^1\) extensions to two questions concerning the minimal constant \(K\) in Fefferman's extensibility test and the number of Glaeser \(K\)-refinements of a standard \(m\)-bundle needed to arrive at a \(K\)-steady bundle or a \(K\)-nonrefinable bundle. A reduction to homogeneous bundles and a more convenient description of Glaeser refinements for standard \(1\)-bundles are obtained. The main result is a construction of a set \(E \subset \mathbb{R}^n\) such that a special bundle, related to \(C^1\) extensions, needs an almost maximal number of refinements until it stabilizes. A new proof of Glaeser's extension theorem using this circle of ideas is also presented.