Nonexpanding attractors: conjugacy to algebraic models and classification in 3-manifolds (Q609747)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Nonexpanding attractors: conjugacy to algebraic models and classification in 3-manifolds |
scientific article |
Statements
Nonexpanding attractors: conjugacy to algebraic models and classification in 3-manifolds (English)
0 references
1 December 2010
0 references
In this very interesting paper, the author states and proves a relevant result in a major theme of dynamical systems. The main result (Theorem 1.1) states that topological mixing, uniform hyperbolicity, and codimension one, of an attractor \(\Lambda\) of a \(C^r\) embedding \(f\), implies highly structured algebraic and topologic restrictions of the subspace \(\Lambda\), and of \(f|_\Lambda\). In fact, the thesis of Theorem 1.1 states the following dichotomy: Either \(\Lambda\) is expanding, i.e., \(\dim\Lambda = \dim(E^u|_\Lambda)\), and \(f|_\Lambda\) is conjugate to the shift map, or \(\Lambda\) is homeomorphic to a compact abelian group (a toral solenoid) and \(f|_\Lambda\) is conjugate to a group automorphism. This result is a non-trivial generalization of the Franks-Newhouse Theorem about Anosov diffeomorphisms [\textit{J. Franks}, in: S.-S. Chern and S. Smale (eds.), Global analysis. Proceedings of the symposium in pure mathematics of the American Mathematical Society. Providence, R.I.: American Mathematical Society (AMS) (1970; Zbl 0207.54304); \textit{S. E. Newhouse}, Am. J. Math. 92, 761--770 (1970; Zbl 0204.56901)], and an extension to proper subsystems of \textit{R. F. Williams}'s results about expanding maps [Publ. Math., Inst. Hautes Étud. Sci. 43, 169-203 (1973; Zbl 0279.58013)]. Theorem 1.1 and its two direct corollaries (Corollary 1.2 and Theorem 1.3) are of great relevance, since proper invariant subsets \(\Lambda\) mostly exhibit dynamics which are not conjugate to fully structured algebraic systems. In Corollary 1.2, the author proves that 2-dimensional basic sets, embedded on 3-dimensional manifolds, are either codimension-1 expanding attractors (or contracting repellers) or decomposable as a finite disjoint union of hyperbolic automorphisms in \(T^2\). As a consequence, in Theorem 1.3, he classifies all the possible hyperbolic attractors in 3-dimensional manifolds, proving that there does not exist more that the classical examples. This results answers a question posed by Bonatti in 1999. The proof of Theorem 1.1, from which the other two results follow easily, takes almost all the paper. It is exact, economical, clear, and principally very beautiful. Along the proof, the author combines dynamics with topology and algebra, showing the very rich and strong mathematical dialog relating them in the abstract case he is studying. He obtains, from the dynamical properties of \(f\), very strong algebraic restrictions via an optimally chosen metrization. Although many mathematical concepts and techniques are combined, the author takes care of precisely defining each of them. Thus, the proof and the whole article are self-contained and very agreeable to read. My rough sketch of the proof of Theorem 1.1 is the following: The author constructs a new metric \(d_{\Omega}\) on \(\widetilde \Omega = \widetilde \Lambda/ \sim \), where \(\widetilde \Lambda\) is the lifting of the hyperbolic attractor \(\Lambda \) to a universal covering space \(\widetilde B\) of its basin of attraction \(B\) and \( \sim \) is an adequate equivalence relation. More precisely, \( \sim \) identifies points in the global stable manifolds contained in maximal (not only local) rectangles of the local product structure in \(\widetilde \Lambda\). In other words, \(\widetilde \Omega\) is intended to be the set of stable manifolds in the lifted space, that is correspondent, after a global holonomy, to each unstable manifold. The metrization \(d_{\Omega}\) induces, in the quotient space, a weaker topology than the one inherited from \(\widetilde \Lambda\) by \( \sim \). To define \(d_{\Omega}\), the author achieves an optimal profit of the following two, respectively topological and measurable, dynamical properties: First, he extends, as much as possible in \(\widetilde \Lambda\), the local product rectangles which exist since \(\Lambda\) is an hyperbolic attractor, and thus, locally maximal invariant. Secondly he considers an optimal \(\sigma\)-compact measure along the lifted unstable manifolds. To do so, he adapts Ruelle-Sullivan's Theorem [\textit{D. Ruelle} and \textit{D. Sullivan}, Topology 14, 319--327 (1975; Zbl 0321.58019)], stating the properties of the canonical disintegration of the Bowen's probability \(\mu\) which maximizes the entropy. This disintegration provides conditional measures \(\mu^u\), along the lifted unstable manifolds, that have the following key properties: They are positive on nonempty open subsets in the internal topologies of the unstable manifolds, and \(f\) expands \(\mu^u\) with exponential rate equal to the topological entropy \(h\). The latter two properties allow the author to define \(d_{\Omega}\), and relate the action of the lifting of \(f\) on the unstable leaves with the topological entropy. After constructing such a well-chosen metric \(d_{\Omega}\), the author considers the first fundamental group \(G\) on the basin of attraction \(B\), and proves that it induces isometric actions on \((\widetilde \Omega, d_{\Omega})\). This result on the isometric action of \(G\) (Corollary 5.17), provides the click idea, which turns the proof of Theorem 1.1 to its end. In fact, the dichotomy in the thesis is obtained after the following discussion in Paragraphs 5.2 and 5.3: Either the fundamental group \(G\) acts properly discontinuously on \(\widetilde \Omega\), or it does not. If it does, the author concludes that \(\Lambda\) is expanding, and the results of Williams are applicable. If it does not, he had constructed a fully global product structure, and thus, he can adapt the proof of \textit{K. Hiraide} of the Franks-Newhouse Theorem [Ergodic Theory Dyn. Syst. 21, No. 3, 801--806 (2001; Zbl 1038.37024)], via the global shadowing lemma.
0 references
uniformly hyperbolic systems
0 references
topological equivalence
0 references
classification
0 references
conjugacy
0 references
3-dimensional dynamical systems
0 references