On Efroymson's extenson theorem for Nash functions (Q1068144): Difference between revisions

Let \(\Omega\) be a semi-algebraic open subset of \({\mathbb{R}}^ n\), and let v be a Nash function on \(\Omega\) and \(V=v^{-1}(0)\neq \emptyset\). The author proves that if \(f_ 1\) is a Nash function defined in a semi- algebraic neighborhood of V, then here exists a Nash function g on \(\Omega\), such that \(f_ 1-g=kv\), where k is a Nash function defined in a semi-algebraic neighborhood of V. This main result of the paper is a little more precise form of a theorem of \textit{G. A. Efroymson} [cf. Géométrie algébrique réelle et formes quadratiques, Journ. S. M. F., Univ. Rennes 1981, Lect. Notes Math. 959, 343-357 (1982; Zbl 0516.14020)]. The author considers his proof as elementary. Probably this depends on the readers background: the author uses the theory of the real spectrum instead of triangulations.