Open conformal systems and perturbations of transfer operators (Q1692404)

The authors of this book focus their attention on the escape rates of conformal open dynamical systems, and the ultimate goal is to understand the asymptotic behavior of the escape rates and the geometry of the survivor sets as the radius of the balls tends to zero. Let \((X, d)\) be a metric space and let \(T : X \to X\) be a continuous map. Consider a ball \(B(z, \epsilon) = \{ x \in X : d(x, z)<\epsilon \}\) centered at \(z \in X\) with radius \(\epsilon>0\). Then we obtain an open system by removing \(B(z, \epsilon)\) from \(X\). Consider for each \(n>0\), the set \(R_n(z, \epsilon)\) of points \(x \in X\) for which \(x, T(x), \dots , T^{n-1}(x) \notin B(z, \epsilon)\), and assume that there exists a \(T\)-invariant probability measure \(\mu\). Then we can consider the measures \(\mu(R_n(z, \epsilon))\) of the sets \(R_n(z, \epsilon)\) as \(n \to \infty\). If the limit of \(\frac{1}{n} \log \mu(R_n(z, \epsilon))\) exists, we call the limit the escape rate from \(B(z, \epsilon)\) and denote it by \(R_\mu(B(z, \epsilon))\). In this book it is considered how the escape rates behave as the radii of the balls tend to zero. The sets \(K_z(\epsilon) = \{ x \in X : T^n(x) \notin B(z, \epsilon) \;\forall \;n \geq 0 \}\) are called the survival sets. Since these have vanishing \(\mu\)-measure, we measure their size and complexity by the Hausdorff dimension HD. The goal is to put our hands on the asymptotic values of \(\text{HD}(K_z(\epsilon))\) when \(\epsilon \searrow 0\) in the context of appropriate conformal dynamics. In the case of hyperbolic conformal systems this has been studied by various authors. The conformal maps considered in this book are very general and their analysis is quite involved. Indeed, the asymptotic existence of escape rates is proved and they are calculated in the context of (finite or infinite) countable alphabets, uniformly contracting conformal graph-directed Markov systems (GDMSs), and in particular, conformal countable alphabet iterated function systems (IFSs). These results have direct applications to interval maps, rational functions and meromorphic maps. Motivated by the work of \textit{A. Ferguson} and the first author of this book [Ergodic Theory Dyn. Syst. 32, No. 3, 961--988 (2012; Zbl 1263.37004)], the authors firstly develop, on a purely symbolic level, an appropriate theory of singular perturbations of Perron-Frobenius transfer operators associated with countable alphabet subshifts of finite type and Hölder continuous summable potentials on countable alphabet subshift of finite type symbol space. This leads to a fairly full account of the structure of the corresponding open dynamical systems and their associated surviving sets. Let us outline the contents of the book. The book consists of five chapters and one appendix. In Chapter 1, the authors give an introduction to the main ideas, techniques and results. Their relations with the relevant previous research on open systems are discussed. In Chapter 2, the authors present the concept and various properties of Perron-Frobenius transfer operators. Using the Keller-Liverani perturbation theorem from Appendix A, the authors deal at length with singular perturbations of Perron-Frobenius operators. The fundamental perturbative theorem (Proposition 2.4.2) enables one to calculate the asymptotic behavior of the leading eigenvalues of perturbed operators. In Chapter 3, the authors define escape rates and prove their existence for open systems generated by open sets \(U_n\). The existence and restricted uniqueness of conditional invariant measures, that are absolutely continuous with respect to the original equilibrium state \(\mu_\varphi\) on \(U_n^c\), are proved. The majority of this chapter is devoted to the study of the shift-invariant survivor set \(K(U_n)\) and to the symbolic dynamics that it generates. For all \(n \geq 1\) large enough, a variational principle for the dynamical system \(\sigma: K(U_n) \to K(U_n)\) is proven. This is not trivial for at least two reasons: first, the sets \(K(U_n)\) need not to be compact; second, it is not known neither whether the subshifts \(\sigma: K(U_n) \to K(U_n)\) are of finite type, nor whether these are topologically mixing. Finally, the authors prove some strong stochastic properties of this dynamical system, such as exponential decay of correlations and an almost sure invariance principle. In Chapter 4, the survivor set and the escape rates along with their asymptotics in the case of conformal countable alphabet (strictly contracting) GDMSs and IFSs are studied. The authors prove the asymptotic existence of escape rates generated by Euclidean balls. From these results the authors deduce their analogues for parabolic GDMSs and IFSs. In Chapter 5, the authors apply their previous results to study the escape rates for multimodal maps of an interval and one-dimensional complex dynamics. The former includes topological Collet-Eckmann maps and subexpanding maps while the latter includes all rational functions of the Riemann sphere and a large class of transcendental meromorphic functions on the complex plane. In Appendix A, the authors formulate and prove the Keller-Liverani perturbation theorem [\textit{G. Keller} and \textit{C. Liverani}, Ann. Sc. Norm. Super. Pisa, Cl. Sci., IV. Ser. 28, No. 1, 141--152 (1999; Zbl 0956.37003)].