A computational model for algebraic power series (Q1184020)

The authors develop a computational model for algebraic formal power series based on a symbolic codification of the series by means of the implicit function theorem. They then reduce the problem of handling a finite set of algebraic series to some corresponding problem involving suitable polynomial rings. In this model they show that most of the usual local commutative algebra can be effectively performed on algebraic series. --- The main result of the paper is an effective version of the Weierstrass preparation theorem: the authors are able to prepare a distinguished polynomial and contemporaneously reduce the involved locally smooth system to one with one less variable. This theorem will allow them to have an effective version of the Weierstrass division theorem, to handle an effective elimination theory for algebraic series and to give an effective version of the Noether normalization lemma. In section 1 the authors recall the basic theory of standard bases in rings of formal power series. The second section is devoted to the presentation of the proposed computational model for algebraic series, based on the concept of locally smooth systems. In section 3 they show how to modify a locally smooth system to compute effectively algebraic series. In section 4 they then give an algorithm to compute a standard basis for the ring of algebraic series. In section 5 they give effective versions of the Weierstrass preparation and division theorems, which are used in section 6 to present algorithms for computing the elimination of variables and the Noether normal position of an ideal of algebraic formal power series. By means of the Artin-Mazur theorem the authors show in the appendix how to reduce classically defined algebraic series to their model and conversely.