Existence of divergent Birkhoff normal forms of Hamiltonian functions (Q372154)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Existence of divergent Birkhoff normal forms of Hamiltonian functions
scientific article

    Statements

    Existence of divergent Birkhoff normal forms of Hamiltonian functions (English)
    0 references
    0 references
    14 October 2013
    0 references
    Any analytic map with a non-resonant fixed point can be transformed through a formal canonical transformation into its Birkhoff normal form. Of course, this formal transformation is in general divergent (see, for instance, [\textit{H. Ito}, Comment. Math. Helv, 51, No. 3, 412--461 (1989;Zbl 0686.58021)]). In the present work, the author deals with the divergence of the Birkhoff normal form itself. In the paper, \(\mathbb{R}^4\) is considered as a symplectic space with the standard symplectic \(2\)-form \(\omega = dx_1\wedge dy_1+dx_2 \wedge dy_2\). If a map, analytic around the origin, has the form \[ h(x_1,x_2,y_1,y_2) = \lambda_1 x_1 y_1 + \lambda_2 x_2 y_2 + O(\|(x_1,x_2,y_1,y_2)\|^3) \] and \(\lambda_1,\lambda_2\) are non-resonant, that is, \(k_1 \lambda_1+k_2 \lambda_2 \neq 0\) for all \((k_1,k_2) \in \mathbb{Z}^2\setminus \{(0,0)\}\), then it is possible to find a formal canonical transformation~\(\varphi\), that is, a formal power series, such that \(h \circ \varphi^{-1}\) is in Birkhoff normal form, namely, it is a formal power series in \(x_1 y_1\) and \(x_2 y_2\). It is well known [\textit{C. L. Siegel}, Ann. Math. (2) 42, 806--822 (1941; Zbl 0025.26503)] that, in general, the formal transformation~\(\varphi\) is divergent. The author, in the present work, proves that there exist some non-resonant \(\lambda_1,\lambda_2\) and an analytic function \(h\) of the above form such that its Birkhoff normal form is divergent. The claim also holds for analytic maps of the form \[ \lambda_1 (x_1^2+ y_1^2) + \lambda_2 (x_2^2 y_2^2) + O(\|(x_1,x_2,y_1,y_2)\|^3) \] with appropriate non-resonant \(\lambda_1,\lambda_2\). The proof is based on the study of how the small divisors that appear in the formal map~\(\varphi\) affect the normal form \(h\circ \varphi^{-1}\) and uses Siegel's arguments in [loc. cit.].
    0 references
    Birkhoff normal form
    0 references
    small divisors
    0 references

    Identifiers