Schubert unions in Grassmann varieties (Q2467308): Difference between revisions

Denote by \(G(l,m)\) the Grassmann variety of \(l\)-dimensional subspaces of a fixed \(m\)-dimensional vector space \(V\) over a field \(F\). Then \(G(l,m)\) is embedded by the Plücker embedding into \(\mathbb{P}^{k-1}\) where \( k = { m \choose l}\). The authors consider the following problem: What is the maximal intersection of a linear subspace of a given dimension in \(\mathbb{P}^{k-1}\) with \(G(l,m)\)? where maximal may refer to the number of \(\mathbb{F}_q\) rational points or to the Krull dimension. The main theorem given as theorem 5.1 states that the maximal Krull dimension of a component is computed by a subspace spanned by certain subvarieties called Schubert unions. They state in fact for \(l = 2\) a numerical criterion in terms of the dimension of the subspace and a maximal number which satisfies an arithmetic quotient, stated as corollary 5.2. The paper is organized as follows. In section two, the authors recall the basic definitions of Schubert cycles with respect to a given Flag. For an ordered \(l\)-tuple \(\alpha\) they recall the definition of a cell which belong to a grid which is partially ordered. They recall the definitions of a Schubert cycle in terms of a union of cells and a partial ordered is introduced for the \(l\)-tuples \(\alpha\). For each \(l\)-dimensional subspace \(W\) of \(V\) a matrix \(M_W\) is associated and therefore by computing all the maximal minors of \(M_W\) form the set of Plücker coordinates indexed with respect to the grid. In fact the Schubert cycle is expressed by the well known property of the vanishing of a subset of Pl''ucker coordinates. The Schubert union is simply the union of all the Schubert cycles. In proposition 2.3 it is proven that for a finite number of Schubert cycles their intersection is a Schubert cycle, with grid equal to the intersection of all the grids. The Schubert union is expressed as a linear section of \(G(l,m)\) in terms of subspace defined in definition 2.1 and the number of \(F_q\) rational points is computed precisely in terms of the \(l\)-tuples in the union of all grids associated to the Schubert cycles. They introduce in definition 2.5 the property of a subset being Borel fixed in terms of the partial order introduced. In particular, for Schubert unions the grid is also Borel fixed (proposition 2.6). If we denote by \(F\) the algebraic closure of \(\mathbb{F}_q\) and let \( B \subset GL(m, F)\) be the set of upper triangular matrices with respect to a fixed basis of \(V\) then they show that if \( X = \{ X_{\alpha} \} \) be a set of Plücker coordinates and \(M(X)\) the corresponding grid, then the linear span of \(X\) is stable under \(B\) if and only if \(M(X)\) is Borel fixed. They show in proposition 2.7 that the Schubert unions of dimension \(r\) are the fixed points under the action of the Borel subgroup of \(G(r, \bigwedge^l V) \). In section 2.1, for a subset \(M \subset G(l,m)\) the geometric dual set \(D(M)\) is given in definition 2.8. For a Schubert union \(S_U\) they define the grid dual which is \(S_U'\) in definition 2.12 and prove in theorem 2.14 that \( D(S_U) = S_U'\). In proposition 2.18 for a Schubert union the number of \(\mathbb{F}_q\)- rational points of its grid dual is computed. In section three, the authors restrict to the case of \(l=2\) and \( m \geq 3\). In theorem 3.2, they prove that for each \(m\) there are \(2^{m-1}\) distinct Schubert unions with respect to a fixed flag. They correspond to the set of subsets \(P(M)\) of a set \(M\) with \(m-1\) elements. Section four is devoted to applications to codes. The authors restrict to \(G_{\mathbb{F}_q}(l,m) \) with cardinality \(n\) and associate to the \(n\) points in it a matrix \(M\) and a code \( C(l,m)\). They first compute in proposition 4.1 the weights for \(C(2,m), m \leq 4\) given already by \textit{D. Yu. Nogin} [in: Arithmetic Geometry and Coding theory, Luminy, 1993, de Gruyter, Berlin, 145--154 (1996; Zbl 0865.94032)]. In proposition 4.2 they compute for \(C(2,5) \) the weight \( d_5\).~Motivated by this, they define in definition 4.4 the number \(J_r\) which is the maximum number of points in a Schubert union which span a linear space of codimension \(r\) and they define \(D_r = n - J_r\) for \( r=1, \ldots, k\). The main result in this section is proposition 4.5 which states that for all \(l,m,r,\) \( d_r \leq D_r\). In the last section, section 5, they prove their main theorem, namely theorem 5.1 which states that for all \(K\) the maximal Krull dimension of a component of a linear section of \(G(l,m)\) with a subspace of Plücker space of a fixed dimension \(K-1\) is computed as a subspace spanned by Schubert unions. As a particular case, they give an arithmetic criterion to compute the maximal Krull dimension of a component of a linear section of \(G(2,m)\) with a subspace in corollary 5.2.