Dimension and recurrence in hyperbolic dynamics (Q929281)

The book presents a unified introduction to the study of dimension and recurrence in hyperbolic dynamics which is based on the thermodynamic formalism developed by Sinai, Ruelle and Bowen in the 1970ies. Though already Bowen in 1979 noticed the relation between the topological pressure and the Hausdorff dimension (see Bowen's equation on p. 25 of the book) the deep connection between these quantities within the framework of Carathéodory dimension characteristics was demonstrated in the work of Pesin and was presented in his book: Dimension Theory in Dynamical Systems, Chicago Univ. Press, 1997. Later this and related approaches to the dimension theory were further enhanced both by Pesin and by other researchers and the author of the present book played an important role in this development. The book consists of two introductory sections and of four parts: Dimension theory, two parts on Multifractal analysis, and the part on Hyperbolicity and recurrence. The first part consists of Chapters 3--5 which start with general Moran geometric constructions, proceed to the thermodynamic formalism for symbolic dynamics and then shows how the latter yields the dimension characteristics of the former. In particular, Section 3.3 there describes nonadditive thermodynamic formalism developed by the author of this book. Part II starts with chapter 6 which gives an introduction to the multifractal analysis and describes it for repellers and hyperbolic sets of conformal maps. In Chapter 7 the author introduces the notion of \(u\)-dimension which provides a general concept of multifractal analysis which is used, in particular, in Chapter 8 to show that the sets of points for which Birkhoff's averages do not converge have full topological entropy and full Hausdorff dimension. Part III deals with additional topics in multifractal analysis, in particular, with applications to number theory. Here there is a small problem with the exposition in Section 10.7. First, in the formula for \(K_\alpha\) on p. 188 the factor \(\frac 1n\) is missing in front of the sum. Secondly, while the result of Eggleston is cited correctly so that the formula (10.17) holds true when even some of the frequencies \(\alpha_i\) are zero, in Proposition 10.7.3 all \(\alpha_i\)'s are supposed to be positive (otherwise some of the formulas in the proof do not make sense), so the result of Eggleston does not follow completely from this proposition. At this point a historical excursus seems appropriate. The first nontrivial use of ergodic theory methods in dimension theory is due to Billingsley who applied them, in particular, in order to prove Eggleston's result which appeared also in his book: Ergodic Theory and Information, Wiley, 1965. The proof there contained a gap which concerns exactly the case when some of frequencies \(\alpha_i\) are zero and then Bernoulli measures constructed by such frequencies give zero weight to some cylinder sets which contain uncountably many points, and so these sets cannot be ignored in computation of the Hausdorff dimension while this ergodic theory approach necessarily disregards them. This gap reappeared later in many places, in particular, in the first edition of \textit{K. Falconer}'s book [Fractal geometry: mathematical foundations and applications, Wiley (1997; Zbl 0871.28009)] and it was first noticed only after more than 30 years by \textit{Y. Kifer} [Trans. Am. Math. Soc. 348, No.5, 2003-2038 (1996; Zbl 0874.28009)] (see p. 2012 and Remark 2.5 there) where we explained that zero frequencies require an additional piece of proof based either on the original Eggleston's combinatorial approach or on a shorter perturbation argument. The final Part IV is devoted to the study of interplay between hyperbolicity and recurrence. Here the notions of pointwise dimensions and product structures of hyperbolic invariant measure play an important role. The final Chapter 15 describes the relation between local quantitative recurrence rates on a hyperbolic set and the pointwise dimension of corresponding invariant measures. The book won the Balaguer 2008 award for expository mathematical monographs and, indeed, the book is quite well written, the author do not strive to discuss immediately the most general case so that the reader can both study the essentials of the modern dimension theory and to have an excellent introduction to many original papers whose survey is provided throughout the monograph. This book is a welcome addition to the literature and it will be useful both to seasoned researchers and to young postdocs and PhD students.