Multiplicities associated to graded families of ideals (Q2435287): Difference between revisions

Let \((R,\mathfrak{m})\) be a Noetherian local ring of dimension \(d\). A graded family of \(\mathfrak{m}\)-primary ideals \(\{I_n\}\) is a family of ideals indexed by the natural numbers such that \(I_0=R\) and \(I_n\) is \(\mathfrak{m}\)-primary for \(n>0\) and \(I_nI_m\subset I_{n+m}\) for all \(n\) and \(m\). Let \(\ell_R(N)\) denote the length of an \(R\)-module \(N\). In this paper it is proved that if \(d>0\) and \(R\) is analytically unramified and equicharacteristic with perfect residue field, or if \(d>0\) and \(R\) is regular then the limit \[ \lim_{n\rightarrow\infty}\frac{\ell_R(R/I_n)}{n^d/d!}\;\;(*) \] exists and is equal to \[ \lim_{p\rightarrow\infty}\frac{1}{p^d}\Big(\lim_{k\rightarrow\infty}\frac{\ell_R(R/I_p^k)}{k^d/d!}\Big)\;\;(**). \] Formerly, the equality of the limits given in \((*)\) and \((**)\) had only been proved in certain particular cases in [\textit{L. Ein} et al., Am. J. Math. 125, No. 2, 409--440 (2003; Zbl 1033.14030)], [\textit{M. Mustaţǎ}, J. Algebra 256, No. 1, 229--249 (2002; Zbl 1076.13500)] and [\textit{R. Lazarsfeld} and \textit{M. Mustaţă}, Ann. Sci. Éc. Norm. Supér. (4) 42, No. 5, 783--835 (2009; Zbl 1182.14004)]. The proof for the existence of the limit in~\((*)\) involves reducing to the case of a complete domain and then finding a suitable valuation of the quotient field of \(R\) of rational rank \(d\) that dominates \(R\). This valuation is then used to construct semigroups \(\Gamma^\prime\subset\mathbb{N}^{d+1}\) and \(\Gamma\subset\mathbb{N}^{d+1}\) corresponding to \(R\) and the family \(\{I_n\}\), respectively, in such a way that \[ \ell_R(R/I_n)=\left[\#\left(\Gamma^\prime\cap(\mathbb{N}^d\times\{n\})\right)\right]-\left[\#\left(\Gamma\cap(\mathbb{N}^d\times\{n\})\right)\right]. \] The existence of the limit in~\((*)\) then follows from a result of \textit{A. Okounkov} [Prog. Math. 213, 329--347 (2003; Zbl 1063.22024)], subsequently developed in [Zbl 1182.14004], stating that under appropriate conditions (satisfied by both \(\Gamma^\prime\) and \(\Gamma\) here): \[ \lim_{n\rightarrow\infty}\frac{\#\left(\Gamma^\prime\cap(\mathbb{N}^d\times\{n\})\right)}{n^d}=\text{vol}(\Delta(\Gamma^\prime))\;\;(\mathrm{and:}\;\;\lim_{n\rightarrow\infty}\frac{\#\left(\Gamma\cap(\mathbb{N}^d\times\{n\})\right)}{n^d}=\text{vol}(\Delta(\Gamma))). \] Here \(\Delta(\Gamma^\prime):=\Sigma(\Gamma^\prime)\cap(\mathbb{R}^d\times\{1\})\) and \(\Delta(\Gamma):=\Sigma(\Gamma)\cap(\mathbb{R}^d\times\{1\})\) are so called \textit{Okounkov bodies}, where \(\Sigma(\Gamma^\prime)\) and \(\Sigma(\Gamma)\) are the closed convex cones of \(\Gamma^\prime\) and \(\Gamma\) in \(\mathbb{R}^{d+1}\), respectively. A number of applications are given, including an asymptotic additivity formula for multiplicities of an arbitrary graded family of ideals, and proof that the epsilon multiplicity of \textit{B. Ulrich} and \textit{J. Validashti} [Math. Proc. Camb. Philos. Soc. 151, No. 1, 95--102 (2011; Zbl 1220.13006)], defined as a lim sup, is actually a limit in some new cases.