Random dynamics and its applications (Q1126733)

This is a review paper, describing ergodic aspects of so called random dynamical systems. The formal definition of this object consists of a probability space \((\Omega, {\mathcal B}, \mu)\) together with a \(\mu\)-preserving ergodic invertible map \(\theta:\Omega\to\Omega\) of another measurable space \(({\mathbb X}, {\mathcal B'})\) and of a measurable subset \(X \subset {\mathbb X}\times\Omega\) with fibers \(X^\omega=\{x\in{\mathbb X}:(x,\omega)\in X\}\in{\mathcal B'}\). The dynamics is given by a measurable map \(\tau:X \to X\) which is defined as a skew product transformation \(\tau(x,\omega)=(F_\omega x, \theta \omega)\). The fiber maps \(F_\omega:X^\omega\to X^{\theta\omega}\) with the composition rule \(F_\omega^n=F_{\theta^{n-1}\omega}\circ\dots\circ F_\omega: X^\omega\to X^{\theta^n\omega}\) are called random transformations. The theory of random transformations concerns mainly the action of fiber maps \(F_\omega\) for typical \(\omega\in\Omega\), i.e., except for subsets of \(\Omega\) of zero \(\mu\)-measure. The present paper mainly deals with chaotic spatial behavior of compositions \(F_\omega^n\) for typical \(\omega\). Chaotic behavior discussed here can be described in such terms as the central limit theorem, the law of iterated logarithm, large deviations, etc. It is shown that the properties above hold true for random expanding in average maps, random subshifts of finite type and certain hyperbolic diffeomorphisms. It is worth to notice that only smooth one-to-one maps without singularities are considered. Applications to computations of fractal dimensions of random sets and random walks in random environment are discussed as well.