Stable reductive varieties. I: Affine varieties (Q705113): Difference between revisions

The motivation for the article is to extend the construction of the moduli spaces of polarized abelian varieties \(A_g\) and their compactifications to a non-commutative setting. In particular, the toroidal compactification due to \textit{G. Kempf}, \textit{F. Knudsen}, \textit{D. Mumford} and \textit{B. Saint-Donat} [``Toroidal embeddings. I'', Lect. Notes Math. 339 (1973; Zbl 0271.14017)] has an interpretation as the moduli space of projective stable pairs \((X,D)\), and the authors generalize this notion. Consider a given non-commutative connected reductive algebraic group \(G\) defined over an algebraically closed field \(k\) of characteristic \(0\). Let \(B\) and \(B^-\) be opposite Borel subgroups of \(G\) with unipotent radicals \(U\), \(U^-\). Then \(T=B\cap B^-\) is a maximal torus of \(G\). The corresponding Weyl group is \(N_G(T)/T\) where \(N_G(T)\) is the normalizer of \(T\). There exists a unique automorphism \(\theta\) of \(G\) such that \(\theta(t)=t^{-1}\) for all \(t\in T\). Then \(\theta\) is involutive; it exchanges \(B\) with \(B^-\), stabilizes \(N\) and \(T\) and fixes pointwise \(W\). \(G\times G\) is a connected reductive group; \(B^-\times B\) and \(B\times B^-\) are opposite Borel groups with common torus \(T\times T\). \(\Theta:(g_1,g_2)\mapsto ((\theta(g_2),\theta(g_1))\) is an involutive automorphism of \(G\times G\), and the diagonal subgroups \(\text{diag}~G\), \(\text{diag}~T\), \(\text{diag}~N\) are \(\Theta\)-invariant, and the induced map on \(\text{diag}~W\) is the identity. The group \(G\) comes equipped with two commuting \(G\)-actions by left and right multiplication. The authors define a class of reductive varieties consisting of irreducible varieties with a \(G\times G\)-action. This class contains the group \(G\) and all of its equivariant embeddings, and is closed under (flat) degenerations with reduced and irreducible fibers. Also a ``toroidal'' compactification is constructed. The adjoint of a \(G\times G\)-variety \(X\) is the \(G\times G\)-variety \(\Theta(X)\) with action twisted by \(\Theta\). A \(G\times G\)-variety \(X\) is self adjoint if it admits an involutive automorphism \(\Theta_X\) such that \(\Theta_X(\gamma x)=\Theta(\gamma)\Theta_X(x)\) for all \(\gamma\in G\times G\) and \(x\in X\). This condition plays a fundamental role throughout the article, in the classification of the stable reductive varieties. This article concentrates on affine stable reductive varieties. They are all classified in terms of combinatorial data; these turn out to be complexes of cones in the weight space of \(G\), invariant under the action of the Weyl group \(W\). All their one parameter degenerations are described: They arise as in the toric case from certain reductive varieties for the larger group \(G_m\times G.\) From the classification it follows that stable reductive varieties are in bijective correspondence with stable toric varieties equipped with a compatible \(W\)-action. For the reductive varieties there exists an associative multiplication law, making each of them into an algebraic semigroup. The authors name these reductive semigroups. The authors finally study families of stable reductive varieties (in the same manner as \textit{E. B. Vinberg} [in: Lie groups and Lie algebras: E. B. Dynkin's Seminar, Transl., Ser. 2, Am. Math. Soc. 169, 145--182 (1995; Zbl 0840.20041)] in his construction of the Vinberg family) and, more generally, varieties with reductive group action. Given a reductive variety \(X\), the authors constructs a family of such varieties with general fiber \(X\), and they prove that any (reduced and irreducible) degeneration of \(X\) arises in that way. They construct a fine moduli space for families of \(G\)-subvarieties of a fixed \(G\)-module. This applies to all multiplicity-finite \(G\)-varieties, and is an application and a generalization of the multigraded Hilbert scheme of \textit{M. Haiman} and \textit{B. Sturmfels} [J. Algebr. Geom. 13, No.4, 725--769 (2004; Zbl 1072.14007)]. This result plays a fundamental role in the follow up article of this, where the moduli space of projective pairs \((X,D)\) is constructed. The article is important, and proves a deep application of Lie groups and geometric invariant theory. Also, the generalization of the classification of the abelian varieties is successful.