Defect of an extension, key polynomials and local uniformization (Q527774): Difference between revisions

The reader is assumed to already know the definitions of a valuation (which is sometimes a pseudo-valuation in this paper) of a graded algebra, of a quasi-excellent local domain, what is local uniformization, and some notions of local uniformization defined in a previous paper by the same author. The definitions of the henselization, of the defect of an extension and of the graded algebra associated to a valuation, of a complete and \(1\)-complete sets of key polynomials, are given. In the following, the extensions are assumed to be simple transcendental extensions \(k(x)\) of fields. So their elements can be seen as polynomials. Intuitively, a complete set of key polynomials is a set of polynomials that makes it possible to compute the valuation of any polynomial. It is \(1\)-complete if it makes possible to compute the valuation of any polynomial with valuation in the smallest non-zero isolated subgroup of the valuation group. Such a set always exists. The author proves that in a defectless extension one can find such a set which is either finite or indexed by \(\mathbb{N}\). He shows the links between the defect of an extension and the degrees of some polynomials in a set of key polynomials. He deduces an effective computation of the defect of an extension in several examples. In the first one, he proves that the valuation on \(k\) extends in an unique way to \(k(x)\). In the last section the author extends some results about local uniformization in quasi-excellent equicharacteristic local domains, of a previous paper by himself, for a valuation satisfying some inductive assumption about defect.