A vector-valued almost sure invariance principle for hyperbolic dynamical systems (Q1019086)

From MaRDI portal
Revision as of 21:49, 30 January 2024 by Import240129110113 (talk | contribs) (Added link to MaRDI item.)
scientific article
Language Label Description Also known as
English
A vector-valued almost sure invariance principle for hyperbolic dynamical systems
scientific article

    Statements

    A vector-valued almost sure invariance principle for hyperbolic dynamical systems (English)
    0 references
    0 references
    0 references
    27 May 2009
    0 references
    Let \(M\) be a manifold, and \(f: M\to M\) be a \(C^1\) diffeomorphism with a nontrivial uniformly hyperbolic basic set \(X\subset M\) (meaning mainly that it obeys axiom A and possesses a dense orbit with moreover some maximality). Let also \(Y\) be an equilibrium measure corresponding to a Hölder potential, and \(\varphi: X\to \mathbb{R}^d\) be a mean zero Hölder observable. The main result asserts that, as \(N\to\infty\), \[ \sum^N_{n=1} \varphi_0 f^n= W_N+O(N^{{2d+3\over 4d+ 7}+\varepsilon}), \] almost everywhere for any \(\varepsilon> 0\), \((W_t)\) being some \(d\)-dimensional Brownian motion. This result is extended then to a non-uniformly hyperbolic setting, under the assumption that the return time function (in \(X\)) lies in \(L^p\), for some \(p> 2\). This general result is applied to a planar periodic Lorentz gas, studied by Sinaï, i.e. a 3-dimensional flow \((q_t,v_t)\in (\mathbb{R}^2\setminus E)\times\mathbb{S}^1\), where \(E\) is a periodic array of disjoint convex \(C^3\) regions, such that the time between collisions with \(\partial E\) is uniformly bounded: then there exists a nondegenerate 2-dimensional Brownian motion \((W_t)\) , such that for almost every initial condition, \(q_t= W_t+ O(t^{7/15+\varepsilon})\) as \(t\to\infty\), for any \(\varepsilon> 0\). An almost sure functional law of the iterated logarithm and refinement of it, for example, can also be deduced from the general main result got by the authors. The probabilistic proof proceeds by a careful study and use of a mixing property, and by blocking arguments, to reduce the case of sums of weakly dependent random variables to the martingale case.
    0 references
    almost sure invariance principle
    0 references
    nonuniform hyperbolicity
    0 references
    Lorentz gases
    0 references
    Brownian motion
    0 references
    Young towers
    0 references
    mixing property
    0 references
    sums of weakly independent variables
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references