of I, denoted bycore(I), is defined to be the intersection of all reductions of I.The core of ideals was first studied by Rees and Sally [RS], partly due to itsconnection to the theorem of Brianc¸on and Skoda. Later, Huneke and Swanson[HuS] determined the core of integrally closed ideals in two-dimensional regularlocal rings and showed a close relationship to Lipman’s adjoint ideal. Recently,Corso, Polini and Ulrich [CPU1,2] gave explicit descriptions for the core of certainideals in Cohen-Macaulay local rings, extending the result of [HuS]. In these twopapers, several questions and conjectures were raised which provided motivationfor our work. More recently, Hyry and Smith [HyS] have shown that the core andits properties are closely related to a conjecture of Kawamata on the existence ofsections for numerically effective line bundles which are adjoint to an ample linebundle over a complex smooth algebraic variety, and they generalize the result in[HuS] to arbitrary dimension and more general rings. Nonetheless, there are manyunanswered questions on the nature of the core. One reason is that it is difficult todetermine the core and there are relatively few computed examples.Our focus in this paper is in effective computation of the core with an eye topartially answering some questions raised in [CPU1,2]. A first approach to under-standing the core was given by Rees and Sally. For an ideal Iin a local Noetherianring (R,m)having analytic spread l, one can take lgeneric generators ofIin aringofthe form R[U
Read full abstract