Positivity of line bundles on general blow ups of \(\mathbb{P}^2\) (Q298021)

The author studies a classical question about positivity notions of line bundles on blow-ups of the complex projective plane along finite sets of points. This theory is quite rich, but there are many open challenging problems, like the Nagata conjecture, which seem to be out of reach at the present state of the art. In the paper the author improves some results around the ampleness, global generation and very ampleness of line bundles, for instance comparing with those from [\textit{T. Szemberg} and \textit{H. Tutaj-Gasińska}, Proc. Am. Math. Soc. 130, No. 9, 2515--2524 (2002; Zbl 1067.14009)]. Let us introduce the following notation. We consider only finite sets of points \(p_{1}, \dots,p_{r} \in \mathbb{P}^{2}_{\mathbb{C}}\) in (very) general position and we denote by \(\pi : X_{r} \rightarrow \mathbb{P}^{2}_{\mathbb{C}}\) the blow up of the complex projective plane along \(p_{1}, \dots,p_{r}\) with \(r\geq 1\). We denote by \(H\) the pull-back of a general line and we consider line bundles of the form \(L = dH - \sum_{i=1}^{r} m_{i}p_{i}\), where \(m_{i}\) denotes multiplicity of \(p_{i}\). These line bundles are associated to linear series of curves of degree \(d\) passing through \(p_{i}\) with multiplicity at least \(m_{i} \geq 1\). The first result of the paper can be formulated as follows. Theorem 1. Let \(L = dH - \sum_{i = 1}^{r} m_{i}E_{i}\) be a line bundle on \(X_{r}\) with \(d>0\) and \(m_{1} \geq \dots \geq m_{r} \geq 0\). Then \(L\) is ample if the following conditions hold: {\parindent=6mm \begin{itemize}\item[a)] \(d > m_{1} + m_{2}\), \item[b)] \(2d > m_{1} + \dots + m_{5}\), \item[c)] \(3d > 2m_{1} + m_{2} + \dots + m_{7}\), \item[d)] \(d^{2} \geq \frac{s+3}{s+2}\sum_{i=1}^{2} m_{i}^{2}\) for \(2 \leq s \leq r\). \end{itemize}} In the homogeneous case \(m_{i} = m\) for all \(i\) we have the following result. Theorem 2. Let \(L = dH - \sum_{i=1}^{r}mE_{i}\) be a line bundle on \(X_{r}\) with \(m>0\). Then \(L\) is ample if {\parindent=6mm \begin{itemize}\item[a)] \(d > \frac{95}{32} \, m\), \item[b)] \(d^{2} \geq \left( \frac{3r + 40}{3r + 39} \right) rm^{2}\). \end{itemize}} Another result provides conditions for global generation. Theorem 3. Let \(L = dH - \sum_{i=1}^{r} m_{i}E_{i}\) be a line bundle on \(X_{r}\) and assume that \(r\geq 5\) and \(m_{i} \geq 2\) for all \(i\). Then \(L\) is globally generated if the following conditions hold: {\parindent=6mm \begin{itemize}\item[a)] \(d+3 > m_{1} + m_{2} + 2\), \item[b)] \(2(d+3) > m_{1} + \dots + m_{5} + 5\), \item[c)] \(3(d+3) > 2m_{1} + m_{2} + \dots + m_{7} + 8\), \item[d)] \((d+3)^{2} \geq \frac{s+3}{s+2} \sum_{i=1}^{s} (m_{i}+1)^{2}\) for \(2 \leq s \leq r\). \end{itemize}} The last result, which we would like to report here, is the following. Theorem 4. Let \(L = dH - \sum_{i=1}^{r} mE_{i}\) be a line bundle on \(X_{r}\) and assume that \(r\geq 3\) and \(m\geq 4\). Then \(L\) is very ample if {\parindent=6mm \begin{itemize}\item[a)] \(d \geq 3m\), \item[b)] \((d+3)^{2} \geq \left( \frac{r+3}{r+2} \right) r (m+1)^{2}\). \end{itemize}} It is worth pointing out that the author provides a large number of interesting examples which allow to see that is some cases obtained bounds (conditions) are sharp.