Perturbation methods for Lyapunov exponents (Q4263376)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 1341826
Language Label Description Also known as
default for all languages
No label defined
    English
    Perturbation methods for Lyapunov exponents
    scientific article; zbMATH DE number 1341826

      Statements

      8 November 2000
      0 references
      Lyapunov exponents
      0 references
      Hörmander conditions
      0 references
      dynamical systems
      0 references
      perturbation
      0 references
      stability
      0 references
      0 references
      Perturbation methods for Lyapunov exponents (English)
      0 references
      This paper surveys some research from the 1980s on the Lyapunov exponents of affine dynamical systems in \(R^d\) perturbed by random noise. The basic setup is \(dx = A_0 x dt + \sigma \sum_{k=1}^r A_k x \circ dF_t^k\), with \(A_0,\dots A_r\) constant \(d\times d\) matrices and \(\circ dF_t^k\) a stochastic differential. The nonnegative real parameter \(\sigma\) titrates the perturbation. The first section quickly summarizes the Oseledets theorem and non-degeneracy conditions, and reproduces a representation of the top Lyapunov exponent, in terms of the solution to a Fokker-Planck equation, due to Arnold, Oeljeklaus, et al. Section 2 then outlines the basic perturbation methods, obtaining an approximation to the top Lyapunov exponent \(\lambda(\sigma)\) for small values of \(\sigma\). These methods are applied in Section 3 to the case of white noise and real noise, to give an explicit expansion of \(\lambda(\sigma)\) in terms of the spectrum of \(A_0\). There is also a comparison of the real-noise and white-noise cases, which is summarized by the statement that ``white noise has a stronger impact than the corresponding real noise with respect to both destabilizing and stabilizing''. Section 4 goes to the other extreme, considering what happens as \(\sigma\to\infty\). A result of Pardoux and Wihstutz is cited, according to which, in the white-noise case, \(\lambda(\sigma)\) goes to infinity like \(\sigma^2\), unless \(A_1,\dots,A_r\) are simultaneously skew-symmetric, in which case it converges to \(1/d\) times the trace of \(A_0\). The latter circumstance is described as ``stabilizing'' noise, and a further result gives necessary and sufficient conditions, in the case \(r=1\), for the noise to be stabilizing. Section 5 concludes the paper with several open problems. For the entire collection see [Zbl 0916.00011].
      0 references

      Identifiers