Response solutions for degenerate reversible harmonic oscillators with zero-average perturbation (Q6063339)

From MaRDI portal
scientific article; zbMATH DE number 7762016
Language Label Description Also known as
English
Response solutions for degenerate reversible harmonic oscillators with zero-average perturbation
scientific article; zbMATH DE number 7762016

    Statements

    Response solutions for degenerate reversible harmonic oscillators with zero-average perturbation (English)
    0 references
    0 references
    0 references
    0 references
    7 November 2023
    0 references
    The authors analyse the existence of response solutions for a class of quasi-periodically forced reversible systems in the form \[ \dot{\theta} = \omega, \qquad \dot{x} = y + \varepsilon f_1 (\theta, x, y), \qquad \dot{y} = \lambda x^l + \varepsilon f_2 (\theta, x, y).\tag{1} \] Note that \(\lambda \ne 0\) and \(l \ge 2\), \(l \in \mathbb{N}\). The existence of response solutions of (1) is also obtained in [\textit{W. Si}, Discr. Cont. Dynam. Syst. 41, No. 8, 3951--3972 (2021; \url{doi:10.3934/dcds.2021023})] if \([f_2 (\theta, 0, 0)]\) satisfies certain non-zero average conditions, where \([g]\) denotes the average of a continuous function \(g\) on \(\mathbb{T}^d\). The aim of the paper under review is to cover the case when \([f_2 (\theta, 0, 0)] = 0\). More precisely, for \(2p < l\), the authors assume that \[ [f_2 (\theta, 0, 0)] = \left[\frac{\partial f_2 (\theta, 0, 0)}{\partial x} \right] = \left[\frac{\partial^2 f_2 (\theta, 0, 0)}{\partial x^2} \right] = \cdots = \left[\frac{\partial^{p-1} f_2 (\theta, 0, 0)}{\partial x^{p-1}} \right] = 0 \] and that \[\lambda^{-1} \left[\frac{\partial^{p} f_2 (\theta, 0, 0)}{\partial x^{p}} \right] < 0 \text{ if } l - p \text{ is even}, \quad \lambda^{-1} \left[\frac{\partial^{p} f_2 (\theta, 0, 0)}{\partial x^{p}} \right] \ne 0 \text{ if } l - p \text{ is odd}. \] The existence of response solutions of (1) is proved in two different main results, where \(\Lambda > 0\) (hyperbolic case) and \(\Lambda < 0\) (elliptic case) are discussed for \[ \Lambda := l \lambda \left(- \lambda^{-1} \left[\frac{\partial^{p} f_2 (\theta, 0, 0)}{\partial x^{p}} \right] \right)^{\frac{l-1}{l-p}} + p \left[\frac{\partial^{p} f_2 (\theta, 0, 0)}{\partial x^{p}} \right] \left(- \lambda^{-1} \left[\frac{\partial^{p} f_2 (\theta, 0, 0)}{\partial x^{p}} \right] \right)^{\frac{p-1}{l-p}}. \] The main results are proved using KAM iterations. In the remaining case \[\lambda^{-1} \left[\frac{\partial^{p} f_2 (\theta, 0, 0)}{\partial x^{p}} \right] > 0, \quad l - p \text{ is even}, \] the authors show that the treated system admits no response solution in most regions.
    0 references
    0 references
    0 references
    0 references
    0 references
    harmonic oscillator
    0 references
    zero-average perturbation
    0 references
    KAM iteration
    0 references
    reversible system
    0 references
    response solution
    0 references
    0 references
    0 references