Periodic orbits in complex Abel equations (Q859530): Difference between revisions

This paper deals with Abel differential equations of the form: \[ \frac{dz}{dt} = z^3 + B(t) z^2 + C(t) z + D(t), \tag{1} \] where \(t \in \mathbb{R}\), \(z \in \mathbb{C}\) and \(B\), \(C\) and \(D\) are smooth, \(2\pi\)-periodic complex valued functions. Two properties of these equations related with its \(2 \pi\)-periodic solutions are shown. The first result establishes that given any \(\ell \in \mathbb{N}\), there is an Abel equation of the form (1) with \(B\), \(C\) and \(D\) trigonometric polynomials, having at least \(\ell\) limit cycles. Therefore, there is no upper bound for the number of isolated \(2 \pi\)-periodic orbits of equation (1). This property is unexpected from the results concerning real Abel differential equations of the form (1); results which are exhaustively described and referred in the introduction of the paper. In strong relation with this result, it is also established that given any \(\ell \in \mathbb{N}\), there are Abel equations of the form \[ \dot{z} = A(t) z^3 + z^2 + C(t) z, \] with \(t \in \mathbb{R}\), \(z \in \mathbb{C}\) and \(A\) and \(C\), \(2 \pi\)-periodic complex valued trigonometric polynomials, having at least \(\ell\) limit cycles. The motivation which leads to the proof of these results consists in the study of how many periodic orbits appear from perturbations of the integrable complex equation \(\dot{z} = z^n(z-i/n)\), where \(i=\sqrt{-1}\), \(n\in \mathbb{N}\) and \(n \geq 2\). First-order perturbative terms are wisely chosen in order to, on one hand, allow the standard study of which and how many periodic orbits persist from a continuum of periodic orbits and, on the other hand, relate the perturbed equation with Abel differential equations. The main theorem summarizing this study states that for each \(k \in \mathbb{N}\) and for each \(\varepsilon\) small enough, there exists a differential equation of the form \[ \frac{dz}{dt} = z \left( z^n - \frac{i}{n} \right) + \varepsilon F(t,z), \] where \(n \geq 2\), \(F(t,z) = \sum_{j=0}^{n} A_j(t) z^j\) with each \(A_j(t)\) a trigonometrical polynomial of degree \(k\), such that the differential equation has at least \(n(k+1)\) hyperbolic \(2 \pi\)-periodic orbits and \(n\) hyperbolic \(2 n \pi\)-periodic orbits. The second main result established in the paper is that the following family of equations of the form (1) \[ \frac{dz}{dt} = z^3 + \left( C_{-1} e^{-ti} + C_0 + C_1 e^{ti} \right) z, \] with \(C_{-1}C_1 \neq 0\), has a center at \(z=0\) if, and only if, \(C_0 = ki\) for some \(k \in \mathbb{Z}\) and \(J_{2| k| }(4i\sqrt{C_{-1}C_1})=0\), where \(J_{2| k| }\) is the Bessel function of order \(2| k| \). We remark that the center variety of this family is formed by infinitely many connected components in the parametric space \(\{ (C_{-1}, C_0, C_1) \in \mathbb{C}^3 : C_{-1}C_1 \neq 0\}\). This example is also quite unexpected if one considers the known results about the center varieties of families of Abel differential equations. The method used to prove this fact involves the explicit expression of the first three Lyapunov quantities associated to the solution \(z=0\) of the equation \(\dot{z} = A(t) z^3 + B(t) z^2 + C(t) z\), which can be encountered in the book [\textit{A. A. Andronov} et al., Qualitative theory of second-order dynamic systems, John Wiley \& Sons, New York (1973; Zbl 0282.34022)]