Livšic theorem for low-dimensional diffeomorphism cocycles (Q268277): Difference between revisions

Let \((M, d)\) be a compact metric space, and let \(f : M \to M\) be a homeomorphism. Given any \(x \in M\) and \(\epsilon> 0\), define the local stable and unstable sets by \[ W^s_\epsilon(x) = \{ y \in M : d(f^n(x), f^n(y)) \leq \epsilon, \forall n \geq 0 \}, \] \[ W^u_\epsilon(x) = \{ y \in M : d(f^n(x), f^n(y)) \leq \epsilon, \forall n \leq 0 \}, \] respectively. We say that a homeomorphism \(f : M \to M\) is hyperbolic with local product structure whenever there exist constants \(\epsilon_0\), \(\delta_0\), \(K_0\), \(\lambda > 0\) and functions \(\nu_s, \nu_u : M \to (0, \infty)\) such that the following conditions are satisfied: \smallskip (h1) \(d(f(y_1), f(y_2)) \leq \nu_s(x) d(y_1, y_2), \forall x \in M, \forall y_1, y_2 \in W^s_{\epsilon_0}(x)\); \smallskip (h2) \(d(f(y_1), f(y_2)) \geq \nu_u(x) d(y_1, y_2), \forall x \in M, \forall y_1, y_2 \in W^u_{\epsilon_0}(x)\); \smallskip (h3) \(\nu_s^{(n)}(x) = \nu_s(f^{n-1}(x)) \cdots \nu_s(x)< K_0 e^{-\lambda n}, \forall x \in M, \forall n \geq 1\); \smallskip (h4) \(\nu_u^{(n)}(x) = \nu_u(f^{n-1}(x)) \cdots \nu_u(x) > K_0 e^{\lambda n}, \forall x \in M, \forall n \geq 1\); \smallskip (h5) If \(d(x, y) \leq \delta_0\), then \(W^u_{\epsilon_0}(x)\) and \(W^s_{\epsilon_0}(y)\) intersect in a unique point which is denoted by \([x, y]\), and it depends continuously on \(x\) and \(y\). \smallskip Notice that shifts of finite type and Axiom A diffeomorphisms (on the non-wandering sets) are particular examples of hyperbolic homeomorphisms with local product structure. The following result is one of the Livšic type theorems obtained in this paper. It is stated that the result, to the best of authors' knowledge, is the first general (i.e., global) result for cocycles taking values in groups of diffeomorphisms of compact manifolds. Let \(f : M \to M\) be a hyperbolic homeomorphism which is transitive and exhibits local product structure, and let \(\Phi : M \to \text{Diff}^1(\mathbb{R}/\mathbb{Z})\) be an \(\alpha\)-Hölder cocycle such that the so called periodic orbit obstructions vanish; i.e., \(\Phi (f^{n-1}(p)) \circ \Phi (f^{n-2}(p)) \circ \cdots \circ \Phi (p) = \mathrm{id}_{\mathbb{R}/\mathbb{Z}}, \forall p \in \text{Fix}(f^n), \forall n \in \mathbb{N}\). Then, there exists an \(\alpha\)-Hölder map \(u : M \to \text{Diff}^1(\mathbb{R}/\mathbb{Z})\) satisfying \(\Phi (x) = u(f(x)) \circ u(x)^{-1}, \forall x \in M\). The authors also obtain a general result (in any dimension) which gives necessary and sufficient conditions to be a coboundary.