Singularities in positive characteristic, stratification and simplification of the singular locus (Q977503): Difference between revisions

Resolution of singularities is one of the central problems in algebraic geometry. \textit{H. Hironaka} proved, in a non-constructive way, that the singularities of a scheme over fields of characteristic zero can be resolved [Ann. Math. (2) 79, 109--203, 205--326 (1964; Zbl 0122.38603)]. Since the late eighties many constructive algorithms have appeared giving an explicit way to resolve singularities, e.g. \textit{O. E. Villamayor U.} [Ann. Sci. Éc. Norm. Supér. (4) 22, No. 1, 1--32 (1989; Zbl 0675.14003)], \textit{E. Bierstone} and \textit{P. D. Milman} [Invent. Math. 128, No. 2, 207-302 (1997; Zbl 0896.14006)], \textit{S. Encinas} and \textit{O. E. Villamayor U.} [Prog. Math. 181, 147--227 (2000; Zbl 0969.14007)]. The algorithms of resolution can be divided into two main steps. The first of them consists in constructing a sequence of monoidal transformations on smooth centers so as to achieve a simplification. This step is done in an inductive way, so as to obtain a monomial ideal supported on the exceptional locus in lower dimension. When this step is finished it is said that the scheme is within the \textit{monomial case}. The second step consists in an easy combinatorial argument that leads to resolution. The first step relies on the existence of hypersurfaces of maximal contact and the existence of some inductive functions that allow to stratify the singular locus on smooth strata. It is well known that the existence of these hypersurfaces fails in positive characteristic [e.g. \textit{R. Narasimhan}, Proc. Am. Math. Soc. 89, 402--406 (1983; Zbl 0554.14010)]. In a previous paper of one of the authors [\textit{O. E. Villamayor U.}, Adv. Math. 213, No. 2, 687--733 (2007; Zbl 1118.14016)], the notion of restriction to hypersurfaces of maximal contact is replaced by transversal projections (elimination of variables) and elimination algebras. The authors prove in this paper, using the elimination approach, a form to extend to arbitrary characteristic the inductive functions (in lower dimension) of the first step. Moreover, they prove that the values of these functions are independent of the chosen transversal projection. This result allow the authors to extend the first step of the algorithms to arbitrary characteristic and define a stratification of the singular locus. As a consequence, they also prove the ``reduction to the monomial case'' in positive characteristic.