Gröbner bases in orders of algebraic number fields. (Q1599544): Difference between revisions

The theory of Gröbner bases is nowadays a well-known tool in the study of polynomials with coefficients in a field. Less known is that Gröbner bases can be computed in the ring of integers [cf. \textit{B. Buchberger}, in: Logic and machines: decision problems and complexity, Proc. Symp., Münster 1983, Lect. Notes Comput. Sci. 171, 137--161 (1984; Zbl 0546.68021)], in the ring of Gaussian integers [\textit{S. Stifter}, J. Symb. Comput. 4, 351--364 (1987; Zbl 0655.13019)], or in a class of rings axiomatically defined by Buchberger [loc. cit.]. A useful extension of this class is due to Stifter [for a different approach, see \textit{K. Madlener} and \textit{B. Reinert} in: Gröbner bases and applications, Lond. Math. Soc. Lect. Note Ser. 251, 408--420 (1998; Zbl 0924.16002)]. In the paper under review it is shown that orders of algebraic number fields are reduction rings in the sense of the papers cited before. The exposition starts by defining a well-ordering, continues by introducing a division algorithm, then demonstrating how to compute a Gröbner basis from any given finite generating set of an ideal. A key point of the approach is a subtle restriction of the set of elements used as multipliers during the algorithm. As a consequence, not every singleton will be a Gröbner basis as in the most familiar case. Used in conjunction with known transfer properties, the theory developed here allows one to effectively perform ideal operations in various rings naturally constructed from orders of algebraic number fields. The paper ends with questions pertaining to future work.