Multiplier ideals of sums via cellular resolutions

From MaRDI portal



Abstract: Fix nonzero ideal sheaves a_1,...,a_r on a normal Q-Gorenstein complex variety X. Fix any positive real number c, and consider the multiplier ideal J of the sum a_1+...+a_r with weighting coefficient c. We construct an exact sequence resolving J by sheaves over X that are direct sums of multiplier ideals for products a_1^{v_1}...a_r^{v_r} for various real vectors v such that v_1+...+v_r = c. The resolution is cellular, in the sense that its boundary maps are encoded by the algebraic chain complex of a regular CW-complex. The CW-complex is naturally expressed as a triangulation T of the simplex of nonnegative real vectors summing to c. The acyclicity of our resolution reduces to that of a cellular free resolution, supported on T, of a related monomial ideal. This acyclicity rests on a comparison between the homology of certain homology-manifolds-with-boundary and the homology of the simplicial complexes obtained by deleting collections of boundary faces from them. Our resolution implies the multiplier ideal sum formula J((a_1+...+a_r)^c) = sum_{|v|=c} J(a_1^{v_1}...a_r^{v_r}), which implicitly follows from Takagi's proof of the two-summand formula (math.AG/0410612). We recover Howald's multiplier ideal formula for monomial ideals (math.AG/0003232) as a special case. Our resolution also yields a new exactness proof for the Skoda complex.


Suppose that \(X\) is a normal \(\mathbb{Q}\)-Gorenstein variety and that \(\mathfrak{a_1}, \dots, \mathfrak{a_n}, \mathfrak{b}\) are ideal sheaves on \(X\). The authors study the multiplier ideal of these ideal sheaves with various exponents. In particular, the authors show that the multiplier ideal \[ \mathcal{J}((\mathfrak{a_1} + \dots + \mathfrak{a_r})^{\alpha} \mathfrak{b}^{\beta}) \] can be resolved by sheaves which are finite direct sums of multiplier ideals \(\mathcal{J}((\mathfrak{a_1}^{\lambda_1} \dots \mathfrak{a_r}^{\lambda_r} \mathfrak{b}^{\beta})\) for various \(\lambda_i \in \mathbb{R}_{\geq 0}\) such that \(\sum_{i = 1}^{r} \lambda_i = \alpha\) (and every such multiplier ideal sheaf appears as a part of the resolution). As a corollary, the authors obtain the following summation formula which generalizes results of \textit{J.-R Demailly}, \textit{L. Ein} and \textit{R. Lazarsfeld} [Mich. Math. J. 48, Spec. Vol., 137--156 (2000; Zbl 1077.14516)], \textit{M. Mustata} [Trans. Am. Math. Soc. 354, No. 1, 205--217 (2002; Zbl 1015.14003)] and \textit{S. Takagi} [Am. J. Math. 128, No. 6, 1345--1362 (2006; Zbl 1109.14005)]. They show that: \[ \mathcal{J}((\mathfrak{a_1} + \dots + \mathfrak{a_r})^{\alpha} \mathfrak{b}^{\beta}) = \sum_{\lambda_1 + \dots+ \lambda_r = \alpha} \mathcal{J}((\mathfrak{a_1}^{\lambda_1} \dots \mathfrak{a_r}^{\lambda_r} \mathfrak{b}^{\beta}). \] The authors also are able to obtain \textit{J. A. Howald}'s characterization [Trans. Am. Math. Soc. 353, No. 7, 2665--2671 (2001; Zbl 0979.13026)] on the multiplier ideal of monomial ideals as a special case. Finally, they also use this technique to provide a new proof of the exactness of the Skoda complex.











This page was built for publication: Multiplier ideals of sums via cellular resolutions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q935878)