Growth of multiplicities of graded families of ideals (Q5964504)

Let \((R,\mathfrak{m})\) be a Noetherian local ring of dimension \(d>0\). Let \(I_\bullet = \{I_n\}_{n\in \mathbb{N}}\) be a graded family of \(\mathfrak{m}\)-primary ideals of \(R\), that is, \(I_0 = R\) and \(I_nI_m \subseteq I_{n+m}\) for all \(n,m \in \mathbb{N}\). The study of the volume \[ \displaystyle \text{vol}(I_\bullet) = \limsup_n \frac{\ell_R(R/I_n)}{n^d/d!} \] of \(I_\bullet\) has attracted the attention of many researchers over the years. The most classical example of such an object is given by choosing \(I_n=I^n\) for all \(n\), in which case one obtains the Hilbert-Samuel multiplicity of an \(\mathfrak{m}\)-primary ideal \(I\). Starting with \textit{A. Okounkov} and his interpretation of asymptotic multiplicities as volumes of certain cones (the Okounkov bodies) [Invent. Math. 125, No. 3, 405--411 (1996; Zbl 0893.52004); Prog. Math. 213, 329--347 (2003; Zbl 1063.22024)], these methods led to an approach used by \textit{R. Lazarsfeld} and \textit{M. Mustaţă} [Ann. Sci. Éc. Norm. Supér. (4) 42, No. 5, 783--835 (2009; Zbl 1182.14004)] and \textit{K. Kaveh} and \textit{A. G. Khovanskii} [Ann. Math. (2) 176, No. 2, 925--978 (2012; Zbl 1270.14022)], who studied graded linear systems on projective schemes. In a series of papers, \textit{S. D. Cutkosky} used different methods to study these objects and, among other achievements, his work shed light on existence of volumes as actual limits [Algebra Number Theory 7, No. 9, 2059--2083 (2013; Zbl 1315.13040)]. In [J. Algebra 442, 260--298 (2015; Zbl 1408.13063)], \textit{S. D. Cutkosky} showed that when \(R\) is a regular local ring and \(I_{n+1} \subseteq I_n\) for all \(n\) (that is, \(I_\bullet\) is a graded filtration), there exists a constant \(\gamma>0\), independent of \(n\), such that \(0 \leq \ell_R(R/I_{n+1})-\ell_R(R/I_n) < \gamma n^{d-1}\) for all \(n \geq 0\). In a sense, even if the function \(n \mapsto \ell_R(R/I_n)\) is not necessarily a polynomial for \(n \gg 0\), under these assumptions it still has some polynomial-like behavior. The main result contained in this article is a generalization of Cutkosky's result, where the assumptions that \(R\) is regular and that \(\{I_n\}\) is a graded filtration are removed. In this case, there exists \(\gamma>0\) such that \[ \ell_R(R/I_{n+1}) -\ell_R(R/I_n) < \gamma n^{d-1} \] for all \(n \geq 0\). Without the assumption that \(I_\bullet\) is a graded filtration, there is no hope of obtaining a finite lower bound. Furthermore, the constant \(\gamma\) is independent of \(n\), but the authors show in Section 5 that it may depend on the graded family \(I_\bullet\). The tools needed to prove the main theorem are developed in Sections 3 and 4. First, the authors analyze the case when the ideals of the graded family are generated by monomials in a polynomial ring over an Artinian ring. Then, they reduce the original problem to this case, by passing to the associated graded ring with respect to a parameter ideal.