On local rigidity of partially hyperbolic affine \(\mathbb{Z}^k\) actions (Q2420200): Difference between revisions

A linear action of \(\mathbb{Z}^k\) on a torus is a morphism \(\rho: \mathbb{Z}^k \to \mathrm{Aut}(\mathbb{T}^n)\), where \(\mathrm{Aut}(\mathbb{T}^n)\) is the group of continuous automorphisms of the \(n\)-torus \(\mathbb{T}^n\). Affine actions of \(\mathbb{Z}^k\) on the other hand also allow for a translational part along with the linear part. We call the action \(\rho\) locally rigid if there exists a neighborhood \(U\) of \(\rho\) in the space \(\mathcal{A}(k,\mathbb{T}^n)\) of smooth \(\mathbb{Z}^k\) actions on \(\mathbb{T}^n\) such that for every \(\eta \in U\) it holds that \(\eta\) is smoothly conjugate to \(\rho\). An important result of \textit{D. Damjanović} and \textit{A. Katok} [Ann. Math. (2) 172, No. 3, 1805--1858 (2010; Zbl 1209.37017)] states that a linear \(\mathbb{Z}^k\) action is locally rigid if and only if it is of higher rank, which is equivalent to condition (HR) for which we refer to the paper under review. Recall that an action of \(\mathbb{Z}^k\) is called of higher rank if \(k>1\) and it does not factor to a \(\mathbb{Z}^k\) action which is, up to a finite index, generated by a single diffeomorphism. For affine actions, a similar result holds by looking at the linear part of the action \(\rho\). The authors of the paper study what happens if the linear part does have rank-one factors. In this case, the action is no longer locally rigid, but it can still be KAM (Kolmogorov-Arnold-Moser) locally rigid and the authors show this is exactly the case when the rank-one factors are trivial. For a precise statement about KAM local rigidity, we refer to part (ii) of Theorem 2.3., which is the main result of the paper. The proof is given in full detail only for the case of a one-dimensional elliptic factor.