[Paper Review] A conjecture on special linear systems of P^3

Antonio Laface, Luca Ugaglia|ArXiv.org|Sep 8, 2004

Polynomial and algebraic computation5 references3 citations

TL;DR

This paper proposes a conjecture classifying special linear systems of hypersurfaces in $\mathbb{P}^3$ with prescribed base points of given multiplicity, using a cubo-cubic Cremona transformation to reduce systems to a standard form. The key contribution is identifying two distinct classes of special systems: those with a fixed quadric component satisfying a negativity condition on intersection numbers, and those with at least one pair of points whose multiplicity sum exceeds the degree by at least 2, leading to a fixed line component.

ABSTRACT

In this note we consider the behavior of linear systems of P^3 through fat points under a cubo-cubic Cremona transformation. This allows us to produce a class of special systems which we conjecture to be the only ones.

Motivation & Objective

• To classify all special linear systems of hypersurfaces in $\mathbb{P}^3$ with prescribed base points of given multiplicity.
• To understand the structure of these systems by analyzing their behavior under a cubo-cubic Cremona transformation.
• To identify the complete set of conditions under which such systems become special, i.e., when their dimension exceeds the expected dimension.
• To provide a conjectural classification of special systems in terms of fixed components—either a fixed quadric or a fixed line—based on intersection-theoretic and combinatorial criteria.

Proposed method

• The authors use a cubo-cubic Cremona transformation to transform linear systems into a standard form where the degree and multiplicities are minimized while preserving dimension.
• They define a 1-cycle $\Gamma({\mathcal{L}})$ associated with pairs of points whose multiplicity sum exceeds the degree by at least 2, indicating fixed line components.
• The virtual dimension of the system is computed using the Riemann-Roch formula on the blow-up $X$ of $\mathbb{P}^3$ at the base points.
• The transformation rule $\mathrm{Cr}({\mathcal{L}}) = {\mathcal{L}}_{3}(d+k, m_1+k, \ldots, m_4+k, m_5, \ldots, m_r)$ is applied, where $k = 2d - \sum_{i=1}^4 m_i$, to reduce the system.
• The difference in virtual dimensions under the Cremona transformation is computed via the formula $v(\mathrm{Cr}({\mathcal{L}})) - v({\mathcal{L}}) = \sum_{t_{ij} \geq 2} \binom{1+t_{ij}}{3} - \sum_{t_{ij} \leq -2} \binom{1-t_{ij}}{3}$, where $t_{ij} = m_i + m_j - d$.
• The conjecture is built on the assumption that special systems arise only when either a fixed quadric exists with negative intersection $Q({\mathcal{L}}-Q)({\mathcal{L}}-K) < 0$, or the 1-cycle $\Gamma({\mathcal{L}})$ has coefficients $t_{ij} \geq 2$.

Experimental results

Research questions

• RQ1 What are the necessary and sufficient conditions for a linear system of hypersurfaces in $\mathbb{P}^3$ with assigned base points and multiplicities to be special, i.e., to have dimension greater than the expected dimension?
• RQ2 How does the cubo-cubic Cremona transformation affect the virtual and actual dimensions of such linear systems?
• RQ3 Can all special linear systems in $\mathbb{P}^3$ be classified into two distinct families: those with a fixed quadric component and those with a fixed line component arising from high multiplicity pairs?
• RQ4 Under what conditions does the residual system after removing fixed components have dimension equal to its virtual dimension, and when does the correction term $\sum \binom{t_{ij}+1}{3}$ account for the speciality?
• RQ5 What is the role of the Harbourne-Hirschowitz conjecture in validating the existence of fixed quadric components in systems with 10 points?

Key findings

• A linear system ${\mathcal{L}}$ in standard form is special if and only if either (i) there exists a quadric $Q$ such that $Q({\mathcal{L}}-Q)({\mathcal{L}}-K) < 0$, or (ii) at least one coefficient $t_{ij} = m_i + m_j - d$ in the 1-cycle $\Gamma({\mathcal{L}})$ is at least 2.
• For homogeneous systems with $r \geq 8$ and $d \leq 2m-1$, the system is empty, implying no special systems exist in this range under the conjecture.
• When $r = 9$ and $d \geq 2m$, the system is special if and only if $2m \leq d < \left[-1 + \frac{3}{2}\sqrt{2m^2 + 2m}\right]$, which defines a finite range of degrees for which speciality occurs.
• If $r \leq 7$ and $d \geq 2m$, the system cannot be special, according to the conjecture and the Harbourne-Hirschowitz conjecture.
• After applying finitely many Cremona transformations, any system can be reduced to a standard form, and the dimension of the original system is given by $\dim{\mathcal{L}} = v({\mathcal{L}}') + \sum_{t'_{ij} \geq 2} \binom{t'_{ij}+1}{3}$, assuming $h^2({\mathcal{L}}' \otimes {\mathcal{I}}_{\Gamma}) = 0$.
• The conjecture implies that all special systems arise from either a fixed quadric or a fixed line with multiplicity at least 2, and no other configurations are possible.