The structure of the core of ideals

From MaRDI portal



Abstract: The core of an R-ideal I is the intersection of all reductions of I. This object was introduced by D. Rees and J. Sally and later studied by C. Huneke and I. Swanson, who showed in particular its connection to J. Lipman's notion of adjoint of an ideal. Being an a priori infinite intersection of ideals, the core is difficult to describe explicitly. We prove in a broad setting that: mcore(I) is a finite intersection of minimal reductions; mcore(I) is a finite intersection of general minimal reductions; mcore(I) is the contraction to R of a `universal' ideal; mcore(I) behaves well under flat extensions. The proofs are based on general multiplicity estimates for certain modules.


The core of an ideal is the intersection of all of its reductions, thus in general an infinite intersection. It was introduced by \textit{D. Rees} and \textit{J. Sally} [Mich. Math. J. 35, 241-254 (1988; Zbl 0666.13004)] in connection with the Briançon-Skoda theorem. The authors prove in this paper that, under some quite general assumptions on the ring, the core is actually a finite intersection, and that it commutes with flat extensions. The main tools are generic ideals associated to a given ideal, and the notion of multiplicity. This paper greatly generalizes the only previously known such results by \textit{C. Huneke} and \textit{I. Swanson} [Mich. Math. J. 42, 193-208 (1995; Zbl 0829.13014)]. It also includes Eisenbud's proof that the core of a monomial ideal is monomial -- a non-trivial observation in view of the fact that reductions of monomial ideals need not be monomial. The last section of the paper is about the computational aspects, using the results on generic ideals from earlier sections. The authors also list further open questions regarding the core.




Cited in
(33)








This page was built for publication: The structure of the core of ideals

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5949432)