\(K\)-orbit closures on \(G/B\) as universal degeneracy loci for flagged vector bundles splitting as direct sums (Q280127): Difference between revisions

Let \(G\) be a classical complex algebraic group, simple up to finite covers, and let \(K = G^\theta\) be the set of fixed points of an involution \(\theta : G \rightarrow G\). Then \(K\) acts with finitely many orbits on the flag variety \(G/B\), and the \(K\)-orbit closures therein can be regarded as generalizations of the Schubert varieties. In [Transform. Groups 18, No. 2, 557--594 (2013; Zbl 1284.14068)], in the case of some particular symmetric pairs \((G,K)\), the author initiated a study of the \(K\)-orbit closures on \(G/B\) from the perspective of torus-equivariant geometry. In the present paper, this study is extended to all the symmetric pairs \((G,K)\) with \(G\) as above. Using equivariant localization and divided difference operators, the author determines formulas for the \(S\)-equivariant fundamental cohomology classes of these generalized Schubert varieties (where \(S\) is a maximal torus of \(K\) contained in a \(\theta\)-stable maximal torus of \(G\)). These formulas are explicit for the closed \(K\)-orbits, whereas for general \(K\)-orbit closures recursive formulas can be produced starting from the closed \(K\)-orbits in terms of the so-called weak order on \(K\backslash G/B\). Motivated by earlier work of \textit{W. Fulton} [Duke Math. J. 65, No. 3, 381--420 (1992; Zbl 0788.14044)], where Schubert varieties are realized as universal degeneracy loci for maps of flagged vector bundles, the author describes an interpretation of these formulas as formulas representing the classes of particular types of degeneracy loci when evaluated at certain Chern classes. Such degeneracy loci are described explicitly in the case \(G = \mathrm{SL}_{p+q}(\mathbb C)\) and \(K = \mathrm{S}(\mathrm{GL}_p(\mathbb C) \times \mathrm{GL}_q(\mathbb C))\), relative to a rank \(p + q\) vector bundle on a smooth complex variety equipped with a flag of subbundles and a splitting as a direct sum of subbundles of ranks \(p\) and \(q\). Similar explicit descriptions are conjectured for all symmetric pairs in types \(B\) and \(C\).