SAGBI bases and degeneration of spherical varieties to toric varieties (Q558703): Difference between revisions

Let \(X\subset\mathbb{P}(V)\) be a complex projective spherical \(G\)-variety, where \(G\) is a classical group and \(V\) a finite-dimensional \(G\)-module. Generalizing the case of toric varieties, one can associate to \(X\) an integral convex polytope \(\Delta(X)\) such that the Hilbert polynomial of \(X\) is the Ehrhardt polynomial of \(\Delta(X)\). The polytope \(\Delta(X)\) is fibred over the moment polytope of \(X\) with the Gelfand-Tsetlin polytopes as fibres. This polytope was defined by Okounkov based on results of Brion. It is proved that, for \(G=\text{Sp}(2n)\), the variety \(X\) can be degenerated, by a flat deformation, to the toric variety corresponding to the polytope \(\Delta(X)\). Firstly, it is known that any spherical variety can be degenerated to a horospherical variety, i.e., to a variety where the stabilizer of a point in the dense \(G\)-orbit contains a maximal unipotent subgroup. Secondly, the main result of the paper states that the homogeneous coordinate ring of a horospherical variety has a (finite) SAGBI basis. Here the homogeneous ring of \(X\) is embedded into a Laurent polynomial algebra and the SAGBI basis is considered with respect to a natural term order. Moreover, the author shows that the semigroup of initial monomials is the semigroup of integral points in the cone over the polytope \(\Delta(X)\). The proof is based on the result of \textit{A.~Okounkov} [in: Kirillov's seminar on representation theory, Am. Math. Transl., Ser.~2, 181, 231--244 (1998; Zbl 0920.20032)]. More general results on degenerations of spherical varieties to toric varieties may be found in [\textit{V.~Alexeev} and \textit{M.~Brion}, Sel. Math., New Ser. 10, 453--478 (2004; Zbl 1078.14075)].