Special effect varieties in higher dimension (Q816234)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Special effect varieties in higher dimension |
scientific article |
Statements
Special effect varieties in higher dimension (English)
0 references
21 February 2006
0 references
The subject of this work are linear systems in \({\mathbb P}^n\) of type \({\mathcal L} = {\mathcal L}_{n,d}(-\sum _{i=1}^h m_iP_i)\), \(P_i\in {\mathbb P}^n\), \(m_i\in {\mathbb N}\); which are given by hypersurfaces of degree \(d\) passing through \(h\) generic points \(P_i\) with multiplicities \(\geq m_i\). Even for \(n=2\), the dimension of \({\mathcal L}\) is not known in general, but several equivalent conjectures (the first due to B. Segre) state that the dimension is the expected one except in the cases when the linear system contains a multiple fixed component which is a \((-1)\)-curve. In a previous paper, the author proposed yet other two equivalent forms for such conjectures, stating that \({\mathcal L}_{2,d}(-\sum _{i=1}^h m_iP_i)\) is special (i.e. it does not have the expected dimension), if and only if it is ``numerically special'' or ``cohomologically special'', where the main interest for those two concepts is that they could be generalized to \(n\geq 3\), where no general conjecture for the dimension of \({\mathcal L}\) is known. We say that \({\mathcal L}\) is numerically special if it exists an \(\alpha\)-special effect variety \(Y\) for \({\mathcal L}\), i.e. an irreducible variety \(Y\) such that \({\mathcal L}-\alpha Y\) has positive virtual dimension greater than \({\mathcal L}\) (a few more conditions are required if \(\dim Y = n-1\)). We say that \({\mathcal L}\) is cohomologically special, instead, if it exists a \(h^1\)-special effect variety \(Y\), i.e. an irreducible \(Y\) such that \({\mathcal L}- Y\) has positive dimension, \(h^0({\mathcal L}|_Y)=0\) and \(h^1({\mathcal L}|_Y)>h^2({\mathcal L}-Y)\) (an example showing that the two definitions do not coincide is given). The author conjectures that for \({\mathcal L}\) to be special is equivalent to being numerically special and also to being cohomologically special. In the case \(n=2\), \(\alpha\)-special effect varieties and \(h^1\)-special effect varieties are actually \((-1)\)-curves, while for \(n\geq 3\) examples are given using rational normal curves, hypersurfaces or linear spaces, and those covers most known examples of special \({\mathcal L}\)'s. Moreover, the case of linear systems in a product \({\mathbb P}^{n_1}\times {\mathbb P}^{n_2}\times\dots\times {\mathbb P}^{n_t}\) is studied and examples (known and new) are given of systems whose speciality is due to special effect varieties.
0 references
linear systems
0 references
multiple points
0 references
fat poins
0 references
Hilbert function
0 references