Linear subspaces, symbolic powers and Nagata type conjectures (Q2445301)

Let \(k\) be an algebraically closed field, \(L_1,\dots , L_s \subset \mathbb{P}^n_ k\) distinct linear subspaces of dimension \(r\) (\(r\)-planes for short), and let \(I = I (L_1\cup \dots\cup L_s ) = I (L_1)\cap \dots \cap I(L_s)\subset k[\mathbb{P}^n] = k[x_0,\dots , x_n]\) be the ideal generated by all forms that vanish on each \(L_i\). In this situation the \(m\)th symbolic power of \(I\) has a simple form: \(I^{(m)} = I (L_1)^m\cap\dots \cap I(L_s )^m\). The main results of this paper concern upper bounds on Waldschmidt constants. These upper bounds are related to a still open conjecture of Nagata, related to his solution of Hilbert's 14th problem. Given \(s\) distinct points in projective space \(\mathbb{P}^n\) (over an algebraically closed field \(k\)), the Waldschmidt constant of the ideal of the points is the infimum of all ratios \(\deg(F )/m\), where \(F\) is any nonzero homogeneous polynomial vanishing to order at least \(m\) at each of the \(s\) points. The conjecture of Nagata states that the Waldschmidt constant equals \(\sqrt{s}\) for \(s\geq 10\) sufficiently general points or \(\mathbb{P}^2\). It is conjectured that equality holds also for each \(n>2\) for \(s \gg 0\) sufficiently general points. Recently, the reviewer, \textit{B. Harbourne} and \textit{A. Van Tuyl} [Adv. Math. 246, 114--127 (2013; Zbl 1296.13017)] have shown how to generalize the upper bound \(\sqrt[n]{s}\) for ideals of disjoint lines in \(\mathbb{P}^3\) . Here, the authors show how it extends to the general setting of disjoint linear subspaces of arbitrary dimension in \(\mathbb{P}^n\) for all \(n>2\), where by linear subspace they mean a subvariety defined by \(n\) or fewer linear forms. The main result is Theorem 2.5 (Section 2), it gives upper bounds on Waldschmidt constant \(\gamma(I)\) of ideals \(I\) of disjoint linear subspaces of projective space. The authors also state three conjectures. Apart from Proposition 4.1, there is not yet much direct evidence for the conjectures. A first step will be to understand how large \(s\) must be for \(s\) disjoint sufficiently general linear subspaces of dimension \(r\) in \(\mathbb{P}^n\) to satisfy the conjectures (see Proposition B.2.1 and Remark B.2.2 for examples where \(s\) is not large enough). At the end, we find two Appendices A and B. The first contains results involving binomial coefficients needed for the proof of Theorem 2.5. Appendix B includes computations demonstrating various techniques for computing \(\gamma(J_{n,r,s})\) for various specific values of \(n, r, s\).