A curve-detecting formula for projective surfaces (Q2326970)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | A curve-detecting formula for projective surfaces |
scientific article |
Statements
A curve-detecting formula for projective surfaces (English)
0 references
10 October 2019
0 references
It is a subtle problem in the theory of algebraic surfaces to decide whether a curve with given intersection properties exists. For instance, this was a key issue in a previous paper by the author and \textit{J. Keum}, Mich. Math. J. 62, No. 1, 3--37 (2013; Zbl 1273.14080)] where an effective obstruction towards existence was developed, based on a somewhat complicated and non-intuitive formula for intersection numbers. The present paper sheds light on this problem by proving a generalized formula with clear meaning which yet is much simpler. In detail, let \(S\) by a (smooth) projective surface with \(p_g=q=0\), so that the Néron-Severi lattice equals the free part of \(H^2(S, \mathbb Z)\). Fixing a \(\mathbb Q\)-basis \(A_1,\hdots, A_r\) of NS\((S)\) with dual basis \(A_1^*,\hdots,A_r^*\), one can express the intersection number of any two \(\mathbb Q\)-divisors \(E, E'\) as \[ (E.E') = \sum_{i=1}^r (E.A_i)(E'.A_i^*). \] (Note that the paper assumes the divisors \(E, E'\) to be effective, but that's not necessary.) The main application which the author has in mind, concerns a (minimal) resolution \(S\stackrel{f}{\longrightarrow}S_0\) of a \(\mathbb Q\)-homology projective plane \(S_0\), where a convenient \(\mathbb Q\)-basis of \(\mathrm{NS}(S)\) is provided by a hyperplane section together with the exceptional curves, or if \(K_{S_0}\not\equiv 0\), its pull-back \(f^* K_{S_0}\) together with the exceptional curves, say \(A_1,\hdots, A_r\). In the latter case, one obtains \[ (E.E') = aa' K_{S_0}^2 + \sum_{i=1}^r (E.A_i)(E'.A_i^*) \] where \(a\) resp. \(a'\) denote the coefficients of the summand \(f^* K_{S_0}\) in \(E\) resp.\ \(E'\) expressed in the above basis (and these summands can be calculated explicitly). This formula turns out to be quite useful, as illustrated by a number of examples.
0 references
lattice
0 references
\({\mathbb{Q}}\)-homology projective plane
0 references
surface singularity, intersection number
0 references