Using symmetric functions to describe the solution set of a zero dimensional ideal
Consider a polynomial ring \(K[\underline{x}]=K[x_1,\ldots ,x_n]\) with char\((K)=0\). Let \(F\supseteq K\) be algebraically closed. Let \(V_F(J)=\{ \Delta_1,\ldots, \Delta_s\} \subset F^n\) be the variety defined by the zero-dimensional ideal \(J\) of \(K[\underline{x}]\) and \(e_i=\) multiplicity of \(\Delta_i\). For any \(u\in K[\underline{x}]\) define in \(K[T]\) the polynomial \(E_{J,u}(T)=\) monic polynomial of degree \(\sum e_i\) having the values \(u(\Delta_i)\) as zeros with multiplicity \(e_i\). There is a unique decomposition \(E_{J,u}(T)=\prod_{j=1}^t (E_{J,u}^{(j)}(T))^j \) of \(E_{J,u}\) into squarefree relatively prime polynomials \(E_{J,u}^{(j)}(T)\). Using the Stickelberger theorem gives convenient matrix representations for the \(K\)-linear mappings \(g+J\mapsto hg+J\) in the algebra \(K[\underline{x}]/J\) (see \textit{P. Pedersen, M.-F. Roy} and \textit{A. Szpirglas} [Prog. Math. 109, 203-224 (1993; Zbl 0806.14042)] or \textit{K. Yokoyama, M. Noro} and \textit{T. Takeshima} [J. Symb. Comput. 14, 399-417 (1992; Zbl 0757.13013)]) and the Newton identities which recover the coefficients of a polynomial from the sum of the \(k\)th-powers of the roots. Given a nonnegative integer \(r\), and a \(k\in \{1,\ldots,n\}\) the authors find an \(F_{J,u}^{(k,r)}\in K[T]\) such that: Theorem 3.8: If \(\alpha\) is a root in \(F\) of \(E_{J,u}^{(j)}(T)\) and \(\beta_1, \ldots,\beta_j\) are the \(x_k\) coordinates of those \(\Delta_i\) for which \(u(\Delta_i)=\alpha\), then \[ \frac{d^{j-1} F_{J,u}^{(k,r)}}{dT^{j-1}} (\alpha) =\frac{1}{j} \left(\sum_{i=1}^j\beta_i^r \right) \frac{d^jE_{J,u} }{dT^j} (\alpha). \] The generalized shape lemma (see \textit{E. Becker, M. G. Marinari, T. Mora} and \textit{C. Traverso} [The shape of the shape lemma, Proceedings of ISSAC-94, 129-33, ACM Press (1993)] and \textit{M.-E. Alonso, E. Becker, M.-F. Roy} and \textit{T. Wörmann} [Prog. Math. 143, 1-15 (1996; Zbl 0879.14033)]) says that if \(u\) is separating for \(J\), i.e. if \(\Delta \mapsto u(\Delta)\) is injective on \(V_F(J)\), then the radical of \(J\) has a description \(\sqrt{J}=\langle f(u),f'(u)x_1-h_1(u), \ldots,f'(u)x_{n}-h_{n}(u) \rangle\). The authors present a new version of that lemma using the polynomials defined above. Finally they present a necessary and sufficient criterion for an element \(u\) to be separating for \(J\). For the entire collection see [Zbl 0847.00060].
- Straight-line programs in geometric elimination theory
- A modular method to compute the rational univariate representation of zero-dimensional ideals
- On the minimum of a positive polynomial over the standard simplex
- Fast computation of special resultants
- scientific article; zbMATH DE number 1163432 (Why is no real title available?)
- Stickelberger and the Eigenvalue Theorem
- Squarefree decomposition of univariate polynomials depending on a parameter. Application to the integration of parametric rational functions
- Symmetry in the core of a zero-dimensional monomial ideal
- Subresultants and the Shape Lemma
- On the computation of matrices of traces and radicals of ideals
- A generalization of Stickelberger's theorem
This page was built for publication: Using symmetric functions to describe the solution set of a zero dimensional ideal
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4352798)
