Stability with respect to actions of real reductive Lie groups (Q1681835): Difference between revisions

The paper under review considers continuous actions of a non-compact real reductive Lie group \(G\) on a Hausdorff topological space \(\mathcal{M}\). The aim of the authors is to develop a geometric invariant theory for such actions, inspired by well-known results for actions on Kähler manifolds. More precisely, write \(G=K\cdot \exp(\mathfrak{p})\) where \(K\) is a maximal compact subgroup of \(G\). In this general setting, in Section 2 the authors define Kempf-Ness functions as maps \(\Psi: \mathcal{M} \times G\to \mathbb R\) satisfying certain conditions. Such a Kempf-Ness function gives rise to a map \(\mathfrak{F}: \mathcal{M}\to \mathfrak{p}^*\), called gradient map. In Section 5, given a point \(x\in \mathcal{M}\), the authors define when \(x\) is polystable (resp. stable, semi-stable, unstable) in terms of the intersection of the \(G\)-orbit of \(x\) with \(\mathfrak{F}^{-1}(0)\). A feature is that these various notions of stability can be detected by a numerical criteria (see Sections 5 and 6), using the maximal weight function introduced in Section 4. The end of the paper is devoted to examples: Section 7 addresses the classical gradient map for Kähler manifolds extending a construction of \textit{I. Mundet i Riera} [Trans. Am. Math. Soc. 362, No. 10, 5169--5187 (2010; Zbl 1201.53086)], Section 8 addresses the action on the space of measures on \(\mathcal{M}\), and Section 9 exemplifies explicitly the latter when \(\mathcal{M}\) is real projective space.