Quantitative statistical stability and linear response for irrational rotations and diffeomorphisms of the circle (Q2070884)

The authors study the ``statistical stability'' of dynamical systems. By ``statistical stability'' they mean how the statistical features of a dynamical system change under perturbation or modification. Their motivation is to understand the extent to which approximations, perturbations, and uncertainties can effect either the qualitative or the quantitative analysis of the resulting dynamics. More generally the authors study systems on a compact metric space \(X\) with a map \(T_0 : X \rightarrow X\), where iterations of \(T_0\) determine the dynamics. They look at systems on \(X\) defined by \(T_\delta : X \rightarrow X\) (with small \(\delta\)) such that \(T_\delta \rightarrow T_0\) as \(\delta \rightarrow 0\). For each \(\delta\), \(\mu_\delta\) is an invariant Borel probability measure for the system \((X, T_\delta)\). The authors investigate the regularity of the map \(\delta \rightarrow \mu_\delta\). Statistical stability of the system \((X, T_0, \mu_0)\) means that the map \(\delta \rightarrow \mu_\delta\) is continuous at \(\delta = 0\) with respect to the topology of the measure space on which \(\mu_0\) is perturbed. The authors take \(X\) to be the circle and consider diffeomorphisms of the circle, and in particular they begin with rotations of the circle. They prove several results that include the following: irrational rotations of the circle under perturbations that are small in the topology of uniform convergence are statistically stable. For Diophantine rotations with the same kind of perturbations they get Hölder statistical stability, where the Hölder exponent depends on the Diophantine type of the rotation number. The authors also prove the statistical stability of diffeomorphisms of the circle under spatial discretization and numerical truncation.