Geometry and Topology
Speaker: Milena Pabiniak (University of Toronto)
"Lower bounds on Gromov width of coadjoint orbits through the Gelfand-Tsetlin pattern."
Time: 15:30
Room: MC 108
Gromov width of a symplectic manifold M is a supremum of capacities of
balls that can be symplectically embedded into M. The definition was
motivated by the Gromov's Non-Squeezing Theorem which says that maps
preserving symplectic structure form a proper subset of volume preserving
maps.
Let G be a compact connected Lie group G, T its maximal torus, and
$\lambda$ be a point in the chosen positive Weyl chamber.
The group G acts on the dual of its Lie algebra by coadjoint action. The
coadjoint orbit, M, through $\lambda$ is canonically a symplectic manifold.
Therefore we can ask the question of its Gromov width.
In many known cases the width is exactly the minimum over the set of
positive results of pairing $\lambda$ with coroots of G:
$$\min \{ \langle \alpha_j^{\vee},\lambda \rangle; \alpha_j \textrm{ a
coroot, }\langle \alpha_j^{\vee},\lambda \rangle>0\}.$$
For example, this result holds if G is the unitary group and M is a complex
Grassmannian or a complete flag manifold satisfying some additional
integrality conditions.
We use the torus action coming from the Gelfand-Tsetlin system to construct
symplectic embeddings of balls. In this way we prove that the above formula
gives the lower bound for Gromov width of U(n) and SO(n) coadjoint orbits.
In the talk I will describe the Gelfand-Tsetlin system and concentrate
mostly on the case of regular U(n) orbits.