Looking out for stable syzygy bundles (Q944314)

The motivation for this paper comes from the theory of tight closure, in particular from an earlier paper of the main author which gives inclusion and exclusion criteria for tight closure in terms of the maximal and minimal slope of twisted syzygy bundles [\textit{H.~Brenner}, Trans. Am. Math. Soc. 356, No.\ 1, 371--392 (2004; Zbl 1041.13002)]. Under the condition that the syzygy bundle is (semi)stable, these criteria yield a simple numerical characterization of tight closure. That is why we are now ``looking out for stable syzygy bundles''. After a comprehensive introduction, Section~1 presents some restriction theorems due to Mehta and Ramanathan, Flenner, Bogomolov and Langer saying that the restriction of a (semi)stable bundle on a smooth projective variety \(X\) to a complete intersection curve of sufficiently high degree is again semistable. In particular, these can be applied for restrictions from \(\mathbb{P}^N\) to a complete intersection curve. If the restriction of a coherent torsion-free sheaf \(\mathcal{S}\) to a generic projective line \(\mathbb{P}^1 \subseteq \mathbb{P}^N\) is semistable, i.e. of the form \(\mathcal{O}(a)\oplus \ldots \oplus \mathcal{O}(a)\), then \(\mathcal{S}\) is semistable. In Section~4, this fact is used to show that the syzygy bundle of \(d+1\) generic forms of degree \(d\) is semistable. Section~5 gives a semistability criterion for a coherent torsion-free sheaf \(\mathcal{S}\) on \(\mathbb{P}^N\) in terms of the exterior powers \(\bigwedge^r \mathcal{S}\). In Section~6, which is the most technical part of the paper, Klyachko's theory of toric bundles is used to obtain a combinatorial-numerical criterion for a monomial ideal to have a stable syzygy bundle on \(\mathbb{P}^N\). Section~7 presents some interesting corollaries and examples of these results. In particular, Corollary 7.1 says that the syzygy bundle of all monomials of fixed degree in a polynomial ring is semistable. The short proof works in any characteristic. Section~8 presents some partial results about the semistability of the syzygy bundle of generic forms of given degrees. Using results by Bohnhorst, Spindler and Hein, combinatorial criteria for semistability and numerical tight closure results similar to those for monomial ideals are shown. In the Appendix by G. Hein, three interesting theorems about the (semi)stability of the syzygy bundles of generic forms of constant degrees are shown. Each theorem gives instances of morphisms \(\bigoplus^n \mathcal{O}_{\mathbb{P}^N} \rightarrow \mathcal{O}_{\mathbb{P}^N}(d)\) whose kernel is a semistable coherent sheaf. The semistability is shown via restriction to some curve \(C \subseteq \mathbb{P}^N\). The first theorem, Theorem~A.1, implies the semistability criterion Theorem~8.6 and the tight closure criterion Corollary~8.7 in the main paper.