Bifurcation of critical periods from the rigid quadratic isochronous vector field (Q924913)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Bifurcation of critical periods from the rigid quadratic isochronous vector field |
scientific article |
Statements
Bifurcation of critical periods from the rigid quadratic isochronous vector field (English)
0 references
29 May 2008
0 references
This paper studies the maximum number of critical periods that can bifurcate from the period annulus of an isochronous center of a planar autonomous differential system under one parameter analytic perturbations. We recall that a critical period is an isolated zero of the derivative of the period function. Some formulae to study the period function is deduced from the expression of the perturbed system in polar coordinates. These formulae are used to study the maximum number of critical periods which can bifurcate from the quadratic rigid isochronous center. Let \(X_0\) be a planar autonomous vector field having an isochronous center at \(p\) and let \(X_\epsilon\) be a one parameter analytic perturbation of \(X_0\) which maintains \(p\) to be a center. The periods of the periodic orbits surrounding \(p\) of \(X_\epsilon\) can be written as \[ T(\xi,\epsilon) = T_0 +\sum_{l\geq 1} T_l(\xi) \epsilon^l, \] where \(\xi\) is a parameterization of these periodic orbits in a real interval \(\mathcal{I}\). Assume that \(T_1(\xi) \equiv T_2(\xi) \equiv \cdots \equiv T_{l-1}(\xi) \equiv 0\) and \(T_i(\xi)\not\equiv 0\). As defined in the paper, when the equation \(T_l'(\xi)=0\) for \(\xi \in \mathcal{I}\) has exactly \(k\) zeros, all of them being simple, \(k\) critical periods are said to bifurcate, for the system corresponding to \(X_0\) and up to \(l\)-th order in \(\epsilon\), from the periodic orbits of the isochronous center. The authors take the transformation to polar coordinates to write the perturbed system and assume that the periodic orbits which belong to the period annulus of the origin of the unperturbed system never cut the curve of zero angular speed. The first result of the paper, Proposition A, shows some explicit formulae for the functions \(T_l(\xi)\), in this setting of polar coordinates. Even though these formulae can be produced for any value of the index \(l\), their expression becomes more complicated to write and to compute with growing \(l\). Thus, only the expressions for \(T_0\), \(T_1(\xi)\) and \(T_2(\xi)\) are explicitly written. The assumption that the unperturbed system is rigid simplifies these formulae, as described in Corollary B, so the rest of the paper is devoted to rigid systems. We recall that an isochronous center is said to be rigid when its angular speed is constant, and thus it can be normalized to be \(1\). The attention is brought to the rigid quadratic system \[ \dot{x} = -y+xy, \quad \dot{y} = x+y^2. \] When studying the maximum number of critical periods which bifurcate from the periodic orbits of this isochronous center, two aspects of the perturbation terms can be considered: either the degree of the perturbation system in the dependent variables \((x,y)\) (when polynomial perturbations are taken into account) or the order in the perturbation parameter \(\epsilon\). Theorem C in the paper establishes some results under perturbations inside the family of quadratic system and up to any order of \(\epsilon\). On the other hand, Theorem D gives an upper bound for the number of critical periods up to first order in \(\epsilon\) and under polynomial perturbations of any degree \(n\). The proof of these results is based upon the formulae described in Proposition A and Corollary B. A remarkable property is given in the proof of Theorem C, see Proposition 3.7, which shows that taking a particular quadratic perturbation of the rigid quadratic system, the expression of \(T_l(\xi)\), provided that \(T_1(\xi) \equiv T_2(\xi) \equiv \cdots \equiv T_{l-1}(\xi) \equiv 0\), only depends on the coefficients of the perturbation terms of the corresponding order \(\epsilon^l\). Moreover, the expression of \(T_l(\xi)\) is the same but with the parameters in the corresponding order \(\epsilon^l\). Thus, the maximum number of critical points of \(T_l(\xi)\) does not depend on \(l\). The result given in Theorem C reinforces Chicone's Conjecture [\textit{C. Chicone}, in her MR review MR 94h:58072 of \textit{W. A. Coppel} and \textit{L. Gavrilov}, Differ. Integral Equ. 6, No. 6, 1357--1365 (1993; Zbl 0780.34023)] on the bifurcation diagram of critical periods of quadratic isochrones. This bifurcation diagram is fully described in the work of \textit{P. Mardešić, D. Marín} and \textit{J. Villadelprat} [J. Differ. Equations 224, No. 1, 120--171 (2006; Zbl 1092.34020)]. In particular, Theorem C implies that over any analytic curve in the parameters space, parameterized by \(\epsilon\), at most one critical period bifurcates from the periodic orbits of the unperturbed rigid isochronous quadratic center.
0 references
critical periods
0 references
isochronous center
0 references
period function
0 references
bifurcations
0 references