A Probabilistic Perspective on the Hamiltonian Diffeomorphism Group
What is a random Hamiltonian diffeomorphism? We propose an answer by constructing a family of probability measures on the Hamiltonian diffeomorphism group Ham(π)
of a closed symplectic manifold π. In the first half of the talk, we outline the construction and show that the family contains measures of full support. Under this family of measures, invariants such as the Hofer norm, spectral invariants, stable commutator length, and homogeneous quasimorphisms are well-behaved random variables. In the second half, we use these measures to define random walks that probe the large-scale geometry of Ham(π) with respect to the Hofer metric. On complex projective space βππ, the expected Hofer norm of the walk after π steps grows at least like πβΎβΎβ. On the other hand, for a random walk confined to a commutative subgroup, it can grow no faster than πβΎβΎβ. We conclude by showing that the rate is exactly of order πβΎβΎβ for the walk on the commutative subgroup generated by ππ-invariant Hamiltonians on βππ.