No label defined (Q3310844)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 3840272
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | No label defined |
scientific article; zbMATH DE number 3840272 |
Statements
1983
0 references
convexity
0 references
compact Lie groups
0 references
loop group
0 references
symplectic
0 references
moment map
0 references
\textit{B. Kostant} proved [Ann. Sci. Éc. Norm. Supér. (4) 6, 413--455 (1973; Zbl 0293.22019)] that if \(G\) is a compact semi-simple Lie group and \( {M} \) a co-adjoint orbit of \( {G} \), then the image of \( {M} \) under the restriction map \( L(G)^{*} \rightarrow L(T)^{*} \) is a convex polyhedron. (Here \( T \) is a maximal torus of \( G \) and \( L(T) \) its Lie algebra.) It is well-known that the infinite dimensional space \( \Omega G \) of maps from the circle \( S^{1} \) into \( G \) (based at the identity) has many properties which are analogous to those of \( G \) and its co-adjoint orbits. This paper proves a version of Kostant's result when \( M \) is replaced by \( \Omega G \). The restriction map \( M \rightarrow L(T)^{*} \) is replaced by the map \( \mu: \Omega G \rightarrow L(T) \oplus \mathbb{R} \) given by \[ \mu(\mathrm{f})=\left(\frac{1}{2 \pi} \mathrm{pr}_{\mathrm{L}(\mathrm{T})} \int_{0}^{2 \pi} \mathrm{f}^{-1} \frac{\mathrm{df}}{\mathrm{d} \theta} \mathrm{d} \theta, \frac{1}{4 \pi} \int_{0}^{2 \pi}\left\|\mathrm{f}^{-1} \frac{\mathrm{df}}{\mathrm{d} \theta}\right\|^{2} \mathrm{~d} \theta\right) \] where the norm \(\|\|\) and the orthogonal projection \( \mathrm{pr}_{\mathrm{L}(\mathrm{T})} \) are defined using the Killing form. (Note that Kostant's map is just \( \operatorname{pr}_{\mathrm{L}(\mathrm{T})} \), after identifying \( \mathrm{L}(\mathrm{T})^{*} \) with \( \mathrm{L}(\mathrm{T}) \).) Then the main result of this paper is that the image of each component of \( \Omega G \) under the map \( \mu \) is the convex hull of the (discrete) set of homomorphisms \( \mathrm{S}^{1} \rightarrow \mathrm{T} \) lying in that component. This is a ``polyhedral paraboloid'' in \( \mathrm{L}(\mathrm{T}) \oplus \mathbb{R} \). Kostant's theorem can be viewed as a result in symplectic geometry, for \( \mathrm{M} \) carries a natural symplectic form by the Kirillov construction, and the map \( M \rightarrow L(T)^{*} \) is just the moment map for the natural action of T on M. In fact, Kostant's theorem follows from general results of \textit{M. F. Atiyah} [Bull. Lond. Math. Soc. 14, 1--15 (1982; Zbl 0482.58013)] and \textit{V. Guillemin} and \textit{S. Sternberg} [Invent. Math. 67, 491--513 (1982; Zbl 0503.58017)] on torus actions on compact symplectic manifolds. The loop group \( \Omega G \) is also a symplectic manifold and \( \mu \) is the moment map of the action of \( T \times S^{1} \) on \( \Omega G \), where \( T \) acts by conjugation and \( S^{1} \) by rotation. More precise results were proved by Atiyah (loc. cit.).for torus actions on compact Kähler manifolds. He showed that the image under the moment map of the closure of any orbit of the corresponding complex torus action is a convex polyhedron, and can be described precisely. This result can be applied to \( \Omega {G} \) by using the Bruhat decomposition of \( \Omega {G} \) into finite dimensional complex cells \( \mathrm{C}_{\lambda} \). Each \( \mathrm{C}_{\lambda} \) is contained in some finite dimensional complex Grassmannian to which the action of \( \mathrm{T} \times \mathrm{S}^{1} \) extends, and each \( \mathrm{C}_{\lambda} \) contains a point whose orbit under the complex torus \( \mathrm{T}_{\mathbb{C}} \times \mathbb{C}^{*} \) is dense in \( \mathrm{C}_{\lambda} \). So Atiyah's result shows that the image of the closure of \( C_{\lambda} \) is a convex polyhedron, which can be determined precisely. The main result follows easily from this.
0 references