On monomial ideals and their socles (Q781537): Difference between revisions

In this paper, the authors consider the following questions. (i) Assume that for a given (a priori unknown) monomial ideal \(I\) the set \(\partial \mathrm{oc}(I)\) is given, is it always possible to retrieve the ideal \(I\) from it? If not, can we always obtain an ideal that contains \(I\)? What additional structure is needed to retrieve \(I\) exactly? In which cases is the ideal we obtain unique? (ii) Given any set of monomials that form an antichain w.r.t. divisibility order, can one always find a monomial ideal \(I\) such that this set of monomials has the form \(\partial \mathrm{oc}(I)\)? Is the ideal \(I\) unique? If not, can they be characterized in some way? \par In order to answer these questions, the authors describe how to constructively obtain a monomial ideal \(I \subseteq R = K[x_1, \ldots, x_d ]\), for a finite subset \(M \subset [x_1, \ldots,x_d ]\) of monomials, such that the set of monomials in \(\mathrm{Soc}(I )\backslash I\) is precisely \(M\), or such that \(M \subseteq R/I\) is a \(K\)-basis for the the socle of \(R/I\). For a given \(M\) the authors obtain a natural class of monomials ideals \(I\) with this property. This is done solely using the lattice structure of the monoid \([x_1, \ldots, x_d].\) Some duality results using anti-isomorphisms between upsets and downsets of the lattice \((\mathbb{Z}^d, \preceq)\) are proved. The authors define and analyze zero-dimensional monomial ideals of \(R\) of type \(k\), where type 1 are exactly the Artinian Gorenstein ideals, and describe the structure of such ideals that correspond to order-generic antichains in \(\mathbb{Z}^d\).