Wonderful subgroups of reductive groups and spherical systems (Q403050)

This work is part of the proof of the Luna conjecture (which states that wonderful varieties are classified by certain combinatorial objects called spherical systems). It follows from \textit{I. V. Losev} [Duke Math. J. 147, No. 2, 315--343 (2009; Zbl 1175.14035)] that a wonderful \(G\)-variety is uniquely determined (if it exists) up to \(G\)-isomorphism by its spherical system. It remains to prove that every spherical system is geometrically realizable, namely that it is actually associated to a wonderful variety. This paper allows to reduce the existence demonstration to a small family of cases, called primitive. In [\textit{P. Bravi} and \textit{G. Pezzini}, ``Primitive wonderful varieties'', \url{arXiv:1106.3187}], the authors complete the proofs showing that actually exists a corresponding wonderful variety for every primitive spherical system. See [\textit{S. Cupit-Foutou}, ``Wonderful varieties: a geometrical realization'', \url{arXiv:0907.2852}] for another her approach to the Luna conjecture. The authors prove some results relating the stabilizer \(H\) (called a wonderful subgroup of \(G\)) of a point in the open \(G\)-orbit of a wonderful \(G\)-variety \(X\), to the spherical system \(S\) of \(X\). In section 3, the authors extracts information on \(H\) only from combinatorial properties of \(S\); in particular they do not suppose a priori that such a \(H\) exists. However, they do not provide a candidate for \(H\). Beyond the classification, Luna's theory describes many geometric constructions and their combinatorial counterparts. For example, morphisms between wonderful varieties correspond to quotients of spherical systems. In the sections 4-6, the authors analyze the relationship between a spherical system \(S\), certain of its quotients, and the relative wonderful subgroups. The authors prove that under certain circumstances the geometric realizability of \(S\) is a consequence of the geometric realizability of these quotients; they produce also a precise description for \(H\). The author also review some results from \textit{D. Luna} [Publ. Math., Inst. Hautes Étud. Sci. 94, 161--226 (2001; Zbl 1085.14039)], refining in particular those on fiber products of wonderful varieties and the the so-called projective fibrations. The above mentioned constructions lead to the definition of primitive spherical systems, a direct generalization of the notion defined in [Zbl 1085.14039] and subsequent works.