On the topology of relative and geometric orbits for actions of algebraic groups over complete fields (Q2437947): Difference between revisions

The geometry of an affine \(k\)-group acting on a \(k\)-variety plays a central role in various areas of mathematics, such as invariant theory and representation theory. The case of algebraically closed \(k\) is studied in depth, for example, in geometric invariant theory. In practice, however, it is important to allow other fields or rings \(k\), such as the complete fields or the adèle rings. The former case is particularly important for noncommutative harmonic analysis on real or \(p\)-adic Lie groups. More precisely, let \(k\) be a field, complete with respect to a real valuation \(v\) of rank one. Let a smooth affine \(k\)-group \(G\) act on a \(k\)-variety \(X\). For \(v \in V(k)\), one may ask whether the relative orbit \(G(k)v\) is open or closed in the geometric orbit \((Gv)(k)\) relative to the Hausdorff topology (that is, the topology induced by \(v\)). The case of characteristic zero has been studied by Borel, Harish-Chandra et al., whereas the positive characteristic case is substantially more difficult: for example, the stabilizer of \(v\) need not be smooth. In this paper, the authors study these questions for more general \(k\) and treat some adélic cases as well. One important ingredient is to equip a topology on the non-abelian cohomologies \(H^1(k, G)\) (using flat or Galois cohomology). These appear naturally in the comparison between relative and geometric orbits. This paper is carefully written and contains a wealth of information.