Symbolic powers of planar point configurations. II. (Q899572)

In the present paper the authors continue their studies on symbolic powers of ideals associated to configurations of points on the projective plane. Recall that for a homogeneous ideal \(I = \bigoplus_{d \geq 0} I_{d}\) one defines the \textit{initial degree} \(\alpha(I)\) of \(I\) as the least degree \(d\) such that \(I_{d} \neq 0\). Let \(Z = \{P_{1},\dots, P_{s} \} \subset \mathbb{P}^{2}\) be a finite set of distinct points and for \(P_{i}\) we denote by \(I(P_{i})\) the radical ideal containing all forms vanishing at \(P_{i}\). In this setting we define the \(m\)-th symbolic power of \(I = I(Z)\) by \[ I^{(m)} := I(P_{1})^{m} \cap \dots \cap I(P_{s})^{m}. \] Now we define the \textit{initial sequence} for \(I\) as the strictly increasing sequence of integers \[ \alpha(I) < \alpha(I^{(2)}) < \alpha(I^{(3)}) < \dots < \alpha(I^{(m)}) < \dots \] and the \textit{first difference sequence} as \[ \beta_{m}(I) = \alpha(I^{(m+1)}) - \alpha(I^{(m)}) \] with \(m \geq 1\) and \(\beta_{0}(I) = \alpha(I).\) Moreover, we define the \textit{Waldschmidt constant} of \(I\) as \[ \widehat{\alpha(I)} := \lim_{m \rightarrow \infty} \frac{\alpha(I^{(m)})}{m}. \] In [\textit{C. Bocci} and \textit{L. Chiantini}, J. Pure Appl. Algebra 215, No. 1, 89--98 (2011; Zbl 1196.14046)] the authors studied the question to what extent the value \(\beta_{1}(I) = 1\) determines the geometry of the set \(Z\) and it turns out that these constraints are quite strong, namely either \(Z\) is contained in a single line \(l\) and \(\alpha(I) = 1\) or \(Z\) is the set of \({d \choose 2}\) points defined by a \(d\)-star configuration of lines \(d\) lines with with \(\alpha(I) = d-1\). Recall that a configuration of lines on the projective plane is called a \(d\)-star configuration if it is a configuration \(d\) lines in general position (no three lines intersect at one point and all intersection points are double points). Recall also that \(Z \subset \mathbb{P}^{2}\) is a quasi star configuration of degree \(d\) if \(Z\) consists of all intersection points of \(d\) general lines and additionally there is exactly one more point from \(Z\) on each of the lines, moreover these additional points are not collinear. In this paper the authors present the following classification result. Theorem. Let \(Z = \{P_{1}, \dots, P_{s}\}\) be a finite set of points in \(\mathbb{P}^{2}\) and let \(I = I(Z)\) be its radical ideal. If \[ \widehat{\alpha(I)} < \frac{9}{4}, \] then \(Z\) {\parindent=6mm \begin{itemize} \item[i)] is contained in a line (\(\widehat{\alpha(I)} = 1\)) or a conic \((\widehat{\alpha(I)} \leq 2)\) or \item [ii)] is the set of all intersection points of \(4\)-star configuration (\(s = 6\)). \end{itemize}} Moreover, if \(\widehat{\alpha(I^{(m)})} = \frac{9}{4}\) for some \(m\), then \(Z\) is \(3\)-quasi star. From the above classification the authors obtain the following two corollaries. Corollary 1. Let \(I\) be the radical ideal of a finite set of points in \(\mathbb{P}^{2}\). Then {\parindent=6mm \begin{itemize} \item[i)] If there exists \(m \geq 1\) such that \(\beta_{m}(I) = \beta_{m+1}(I)=1\), then \(\alpha(I) = 1\). \item [ii)] If there exists \(m \geq 1\) such that \(\beta_{m}(I) = \beta_{m+1}(I) =\beta_{m+2}(I) = \beta_{m+3}(I) = \beta_{m+4}(I) = 2\), then \(\alpha(I) = 2\). \item [iii)] For any \(d \geq 3\) there exist configurations of points such that \(\beta_{m}(I) = d\) for all \(m \geq 2\) but \(\alpha(I) \geq d+1\). \end{itemize}} Corollary 2. Let \(I\) be the radical idea of a finite set \(Z\) of points on \(\mathbb{P}^{2}\) with an integral Waldschmidt constant \(\widehat{\alpha(I)}=d\). {\parindent=6mm \begin{itemize} \item[i)] If \(d=1\), then \(\alpha(I) = 1\). \item [ii)] If \(d\geq 2\), then \(\alpha(I)\) need not be equal to \(d\). \end{itemize}} One of the ingredients of all presented proofs is the notion of \textit{Bézout decompositions}, which follows directly from properties of Bézout's Theorem.