Multiplicities and Rees valuations (Q847023)

The purpose of this article is to establish connections between the Rees valuations of an ideal \(I\) and various multiplicities of \(I\) such as the \(j\)-multiplicity and the \(\varepsilon\)-multiplicity. Let \((R,\mathfrak m)\) be a Noetherian local ring and \(I\) an ideal of \(R\). It is well known that when \(I\) is \(\mathfrak m\)-primary then \(j(I)=e(I)\), where \(e(I)\) and \(j(I)\) denote the usual and \(j\)-multiplicity of \(I\). The authors are concerned with non \(\mathfrak m\)-primary ideals with maximal analytic spread, namely \(d\). They show that in this case there exist positive integers \(d_i(I,\nu_i)\) such that \(j(I)=\sum_i d_i(I,\nu_i)\cdot\nu_i(I)\), where the \(\nu_i\) are the Rees valuations of \(I\) centered on \(\mathfrak m\) such that each \(\nu_i\) is an \(\mathfrak m\)-valuation. This formula generalizes a multiplicity formula due to Rees in the case of \(\mathfrak m\)-primary ideals. The \(\varepsilon\)-multiplicity of an ideal \(I\) is defined to be \(\varepsilon(I)=\limsup_{n\to\infty}\frac{d!}{n^d}\cdot\lambda(\Gamma_{\mathfrak m}(R/I^n))\), where \(\Gamma_{\mathfrak m}\) is the zero-th local cohomology functor. They show that this multiplicity is invariant up to integral closure. Furthermore, \(\varepsilon(I)\neq 0\) if and only if \(I\) has maximal analytic spread. Finally, in the appendix they discuss the one-to-one correspondence between the Rees valuations of \(I\) centered on \(\mathfrak m\) and the Rees valuations of \(I\widehat R\) centered on \(\mathfrak m\widehat R\), where \(\widehat R\) is the completion of \(R\).