Equations for point configurations to Lie on a rational normal curve (Q1621468): Difference between revisions
From MaRDI portal
Set OpenAlex properties. |
Changed an Item |
||
Property / arXiv ID | |||
Property / arXiv ID: 1711.06286 / rank | |||
Normal rank |
Revision as of 19:55, 18 April 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Equations for point configurations to Lie on a rational normal curve |
scientific article |
Statements
Equations for point configurations to Lie on a rational normal curve (English)
0 references
8 November 2018
0 references
The aim of this paper is to study configurations of points in projective space (over an algebrically closed field) which lie on rational normal curves. Namely, when considering sets of \(n\) ordered points in \({\mathbb P}^d\) that lie on a rational normal curve, we get that they give a subset of \(({\mathbb P}^d)^n\), and by taking its closure in the Zariski topology, we obtain a variety that is called its Veronese compactification: \(V_{d,n} \subset ({\mathbb P}^d)^n\). Any point of \(V_{d,n}\) parameterizes a configuration of \(n\) (possibly non-distinct) points supported on a flat limit of a rational normal curve. What the authors wants is to find equations for \(V_{d,n}\). We know that \(V_{d,n} = ({\mathbb P}^d)^n\) whenever \(d=1\) or \(n\leq d+3\), hence we can assume \(d\geq 2\) and \(n\geq d+4\); \(V_{d,n}\) is shown here to be irreducible and of dimension \(d^2 +2d +n-3\). When \(d=2\), a quite complete description of the equations for \(V_{2,n}\) can be given; namely what is shown is that: \begin{itemize} \item[--] \(V_{2,n}\) is defined scheme-theoretically by \({(\frac{n}{6})}\) determinants of certain \(6\times 6\) matrices of quadratic monomials (hence by degree \(12\) equations). \item[--] Let \(A = \{1,\dots, {(\frac{n}{6})}\}\) be a set of indices for those determinants; a subset of those determinants indixed by \({\mathcal T}\subset A\) defines \(V_{2,n}\) set-theoretically if and only if for all partitions \(I_1\cup \dots \cup I_6 = \{1,\dots,n \}\), it \(\exists {\mathcal J}\subset {\mathcal T}\) which meets every set \(I_j\) of the partition in one element. Consequently, at least \(\frac{2}{n-4}{(\frac{n}{6})}\) of those determinants are needed to define \(V_{2,n}\) set-theoretically. \item[--] \(V_{2,n}\) is Cohen-Macaulay, normal and it is Gorenstein if only if \(n=6\). \end{itemize} The case \(d\geq 3\) reveals to be far more complex; what is shown here is that, via the Gale transform, equations for \(V_{d,d+4}\) can be obtained from the ones of \(V_{2,d+4}\). Such equations, via the forgetful maps, give equations for any \(V_{d,n}\) which define a variety \(W_{d,n} \supseteq V_{d,n}\), and it is conjectured that \(W_{d,n} = V_{d,n}\cup Y_{d,n}\), where \(Y_{d,n}\) is the locus of degenerate points configurations (its equations are known). The inclusion \(W_{d,n} \supseteq V_{d,n}\cup Y_{d,n}\) is proved here for all \(d,n\), while the equality is obtained set-theoretically for \(d=3\) and \(n=d+4\). Other partial results are given for \(d=3,4\) and their study shows how the situation gets quite complex.
0 references
parameter space
0 references
point configuration
0 references
gale transform
0 references
rational normal curve
0 references