The quantitative behavior of asymptotic syzygies for Hirzebruch surfaces (Q2154971): Difference between revisions

In the paper under review, the author studies the so-called Ein, Erman, and Lazarsfeld's normality heuristic [\textit{L. Ein} et al., J. Reine Angew. Math. 702, 55--75 (2015; Zbl 1338.13023)] for the asymptotic linear syzygies of Hirzebruch surfaces embedded by \(\mathcal{O}(d,2)\). More specifically, let \(X\) be projective variety of dimension \(n\) over an arbitrary field \(\mathbb{K}\). Given a sequence of very ample line bundles \(\{L_{d}\}_{d\in \mathbb{N}}\), one wants to study how the graded Betti numbers of \(X\) behave asymptotically with respect to \(L_{d}\) if \(d \gg 0\), that is, one is interested in the syzygies of the section ring \[R(X;L_{d}) :=\bigoplus_{k \in \mathbb{Z}}H^{0}(X,kL_{d})\] as a module over \(S = \mathrm{Sym} \, H^{0}(X,L_{d})\). Considering the graded minimal free resolution \[0 \rightarrow F_{r_{d}} \rightarrow \cdots \quad \cdots \rightarrow F_{1} \rightarrow F_{0} \rightarrow R(X;L_{d}) \rightarrow 0,\] let \[K_{p,q}(X;L_{d}) := \mathrm{span}_{\mathbb{K}} \langle\text{minimal generators of } \, F_{p} \, \text{ of degree } \, (p+q)\rangle\] be the finite dimensional \(\mathbb{K}\)-vector space of minimal syzygies of homological degree \(p\) and degree \(p+q\). We write \(k_{p,q}(X;L_{d})\) for \(\mathrm{dim} \, K_{p,q}(X;L_{d})\), and then form the Betti table of \((X;L_{d})\) by placing \(k_{p,q}(X;L_{d})\) in the \((q,p)\)-th spot. In this setup, we can state Ein, Erman, and Lazarsfeld's heuristic as follows: if \(\{L_{d}\}_{d\in \mathbb{N}}\) is a sequence of line bundles growing in positivity, then for any \(q \in [1, ..., n]\) there exists a function \(F_{q}(d)\), depending on \(X\), such that if \(\{p_{d}\}_{d\in \mathbb{N}}\) is a sequence of non-negative integers such that \[(\star): \quad \quad \mathrm{lim}_{d \rightarrow \infty}\bigg(p_{d} - (r_{d}/2 + a\sqrt{r_{d}}/2)\bigg) = 0,\] where \(a \in \mathbb{R}\) is a fixed constant, then \[F_{q}(d) \cdot k_{p_{d},q}(X;L_{d}) \rightarrow e^{-a^{2}/2}.\] Now we can formulate the main results of the paper. Denote by \(\mathbb{F}_{t}\) the Hirzebruch surface embedded by the line bundle \(\mathcal{O}_{\mathbb{F}_{t}}(d,2)\). Theorem A. If \(\{p_{d}\}_{d \in \mathbb{N}}\) is a sequence of non-negative integers satisfying \((\star)\) for some real number \(a \in \mathbb{R}\), then \[\frac{3\sqrt{2\pi}}{2^{r_{d}}\sqrt{r_{d}}} \cdot k_{p_{d},1}(\mathbb{F}_{t}, \mathcal{O}_{\mathbb{F}_{t}}(d,2)) = e^{-a^{2}/2}\bigg(1 + O\bigg(\frac{1}{\sqrt{r_{d}}}\bigg)\bigg).\] Theorem B. There does not exist a function \(F_{2}(d)\) such that if \(\{p_{d}\}_{d\in \mathbb{N}}\) is a sequence of non-negative integers satisfying \((\star)\) for some real number \(a \in \mathbb{R}\), then \[F_{2}(d)\cdot k_{p_{d},2}(\mathbb{F}_{t},\mathcal{O}_{\mathbb{F}_{t}}(d,2)) = e^{-a^{2}/2}\bigg(1 + O\bigg(\frac{1}{\sqrt{r_{d}}}\bigg)\bigg).\]