On Nash images of Euclidean spaces (Q1647394): Difference between revisions

As we all know, images of fairly regular sets ( real analytic, semi-analytic \dots) by very regular maps can go crazy. The following problem was posed by \textit{J. M. Gamboa} [``Algebraic images of the real plane'', in: Reelle algebraische Geometrie, 10--16 June 1990, Oberwolfach, Germany. Oberwolfach: Mathematisches Forschungsinstitut Oberwolfach (1990)]: which subsets of \(\mathbb R^n\) are images of another Euclidean space \(\mathbb R^m\) by a polynomial mapping? The author of this paper studied \textit{M. Shiota}'s [Nash manifolds. Berlin: Springer-Verlag (1987; Zbl 0629.58002)] variation of this question: which \(d\)-dimensional semialgebraic subsets of \(\mathbb R^n\) are Nash images of \(\mathbb R^d\)? Shiota made a conjecture (see Conjecture 1.3) and worked on it with several Spanish mathematicians (see the bibliography). The proof turned out to be difficult and it took several years and several results to come to the main theorem of this paper (Theorem 1.4), which gives seven necessary and sufficient conditions for a \(d\)-dimensional semialgebraic set to be a Nash image of \(\mathbb R^d\). The condition number (v) is that of Shiota: The set S is pure dimensional and there exists a Nash path whose image meets all the connected components of reg\(S\). The author remarks that an important milestone was the following Theorem 1.5 discussed in Section 6 of the paper: Every connected \(d\)-dimensional Nash submanifold (with boundary) of \(\mathbb R^n\) is a Nash image of \(\mathbb R^d\). Nash functions for beginners are very well presented by M. Coste (see his personal page for details). The proof of Theorem 1.4 uses several tools and is a result of long years of works and discussions (see Introduction). Very different techniques are used, for instance blowing up, symmetry by arcs [\textit{K. Kurdyka}, Math. Ann. 282, No. 3, 445--462 (1988; Zbl 0686.14027)], properties of the set of regular points [\textit{J. Stasica}, Ann. Pol. Math. 82, No. 2, 149--153 (2003; Zbl 1056.14080)], properties of Nash functions (some proven only recently) etc. The bibliography is very thorough and the paper is very well written. As an added value, the author proves two remarkable consequences of the main theorem (see 1.B.1 Representation of arc-symmetric semialgebraic sets, Corollary 1.7 and 1.8).