Distance expanding random mappings, thermodynamical formalism, Gibbs measures and fractal geometry (Q639226)

This book treats random dynamical systems. Let \(\theta\) be an invertible ergodic transformation on a probability space \((X,\mathcal B,m)\), and let \(T_x\) \((x\in X)\) be transformations on fibers \(T_x: \mathcal I_x\to\mathcal I_{\theta(x)}\). The authors investigate dynamical systems of the form \[ T^n_x=T_{\theta^{n-1}(x)}\circ\cdots\circ T_x \] on \(\mathcal I=\bigcup_{x\in X}\{x\}\times\mathcal I_x\). They assume only the mean expandedness such as \(\gamma_x>1\) a.e. or \(\int\log\gamma_x\,dm(x)>0\). First, the authors consider a pressure function \(P_x(\phi)\) with potential function \(\phi\) under general conditions. They establish the random version of the Ruelle-Perron-Frobenius theorem concerning the spectra of the transfer operators. Using this theorem, they study the existence, uniqueness and the ergodicity of the Gibbs measures \(\{\nu_x\}\), and also show the exponential decay of correlation. Next, the fractal structure induced by conformal mappings is studied. Let \(f_x\) be a \(C^{1+\alpha}\) conformal and expanding mapping on a smooth Riemannian manifold, and consider a potential \(\phi_t(x,z)=-t\log|f_x'(z)|\). Consider the expected value of the pressure \[ \mathcal EP(t)=\int P(\phi_t(x,z))\,dm_x(z). \] Then the following version of Bowen's formula holds: ``The zero of the function \(t\mapsto \mathcal EP(t)\) is \(m\)-a.e. equal to the Hausdorff dimension of \(\mathcal I_x\).'' The authors also consider multi-fractal spectrum, random Julia set, and real-analyticity of the pressure.