Determining singular solutions of polynomial systems via symbolic-numeric reduction to geometric involutive forms (Q765856): Difference between revisions

Consider a polynomial system \(F = \{f_1,\dots,f_t\} \subset \mathbb{C}[x_1,\dots, x_s]\) which has an isolated singularity at \(\hat{\mathbf{x}} = (\hat{x}_1, \dots, \hat{x}_s)\). The authors of the paper study the singularity by computing (a) the exponent \(\rho\) in the primary component \(Q=\langle f_1,\dots, f_t, P^\rho \rangle\), where \(P\) is the maximal prime corresponding to~\(\hat{\mathbf{x}}\), (b) the \(\mathbb{C}\)-dimension \(\mu\) of \(\mathbb{C}[\mathbf{x}]/Q\), and (c) a basis of the space of differential operators that annihilates \(\langle f_1, \dots, f_t \rangle\) locally at \(\hat{\mathbf{x}}\) (dubbed the \textit{Max-Noether space}). The paper at hand focusses on precision and complexity questions of the symbolic-numeric procedure introduced earlier by the authors [in: ISSAC 2008, 325--332, ACM, New York (2008; Zbl 1236.68302)]. The unsurprising result is that the algorithm has the same complexity as computing the nullspace of a matrix, bounded in terms of the dimension of the matrix.