Kähler Soliton Surfaces Are Generically Toric

Abstract

Let $(M,g,\omega ,f,\lambda )$ be a Kähler gradient Ricci soliton in real dimension four. One first observes that it is an integrable Hamiltonian system in a classical sense. Indeed, all known complete examples are toric and the symmetry is intrinsically related to the potential function $f$ and the scalar curvature $\SS$. While another article addresses the case that these functions are functionally dependent, this one considers the independent case. The main result states that the soliton admits a toric action under a generic assumption. That is, one assumes that the system is non-degenerate and the potential function $f$ is proper. Then there is an effective, completely integrable Hamiltonian toric ${\mathbb{T}}^2$- action on $(M,\omega )$.

improve the main statement (using a result of Munteanu-Wang to remove an assumption for the steady case); improve the abstract and Introduction to make clearer the connection to another article