Bifurcation of 2-periodic orbits from non-hyperbolic fixed points (Q2405387)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Bifurcation of 2-periodic orbits from non-hyperbolic fixed points
scientific article

    Statements

    Bifurcation of 2-periodic orbits from non-hyperbolic fixed points (English)
    0 references
    0 references
    0 references
    0 references
    25 September 2017
    0 references
    The authors introduce the concept of \(2\)-cyclicity for families of one-dimensional maps with a non-hyperbolic fixed point by analogy to the cyclicity for families of planar vector fields with a weak focus. As it turns out, this new concept is useful in order to study the number of \(2\)-periodic orbits that can bifurcate from the fixed point. As an application, they also study the \(2\)-cyclicity of some natural families of polynomial maps. Let \(\Omega \subset \mathbb{R}^{n}\) be an open set, \(N\in \mathbb{N}\cup \left\{0\right\}\) and consider the family of analytic orientation-reversing maps from \(\mathbb{R}\) into itself, \[ f_{a}(x) = -x + \sum_{i\geq 2}c_{i}(a)x^{i}, \quad a\in \Omega. \] One says that the origin of a map \(f_{a*}\) , with \(a*\in \Omega\), has \(2\)-cyclicity \(N\) if: (i) it is possible to find \(\varepsilon_{0}>0\) and \(\delta_{0}>0\) such that the maximum of isolated \(2\)-periodic orbits within \(D_{\delta_{0}}(0) \subset \mathbb{R}\) for every map for which \(a\in D_{\varepsilon_{0}}(a*)\subset \Omega\) is \(N\). (ii) for any \(\varepsilon> 0\) and any \(\delta>0\) there exists \(a\in D_{\varepsilon}(a*)\subset \Omega\) such that \(f_{a}\) has \(N\) different isolated \(2\)-periodic orbits within \(D_{\delta}(0)\). The main result deals with a special case: the maps \(f_{a}\) are polynomials of fixed degree \(d\), and the parameters are the coefficients of the system (notice that the only involution in these families corresponds to the trivial case \(f_{0}(x) = -x\).) One of the main result in this paper is the following claim (Theorem 2): consider the family of polynomial maps \[ f_{a}(x) = -x + \sum_{i=2}^{d}c_{i}(a)x^{i}, \quad a=(a_{2}, \ldots, a_{d})\in \mathbb{R}^{d-1}. \] It has only the trivial involution corresponding to \(a = 0\) and its cyclicity is at most \(\left[\frac{d^{2}-1}{2}\right]\), where \(\left[ \cdot \right]\) stands for the integer part. Furthermore: (i) for \(d\) even, its cyclicity is at least \(d-2\); (ii) for \(d\) odd, its cyclicity is at least \(d-3\).
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    non-hyperbolic fixed point
    0 references
    periodic point
    0 references
    bifurcation
    0 references
    cyclicity
    0 references
    0 references
    0 references