Converting bases with the Gröbner walk (Q1369804): Difference between revisions

The authors present an algorithm which converts a given Gröbner basis of a polynomial ideal \(I\) to a Gröbner basis of \(I\) with respect to another term-order. The technique proposed uses the notion of Gröbner cone and Gröbner fan introduced by \textit{T. Mora} and \textit{L. Robbiano} [J. Symb. Comput. 6, No. 2/3, 183-208 (1988; Zbl 0668.13017)] and is based on the idea that two term orders giving adjacent Gröbner cones can be viewed as a refinement of a common partial order given by a weight vector \(\omega\); moreover in this case the transformation from one Gröbner basis to the other involves only the initial forms of polynomials with respect to \(\omega\) and can be done with linear algebra techniques. If \(<_1\) and \(<_2\) are not adjacent term orders, one can found a path connecting \(<_1\) to \(<_2\) in the Gröbner fan in such a way that the conversion is done with a finite number of conversions of adjacent Gröbner bases.