Linearizability of the polynomial differential systems with a resonant singular point (Q2472850): Difference between revisions

The authors study linearizability and integrability of the planar real systems \[ \dot x = x + P(x,y), \quad \dot y = -\lambda y + Q(x,y), \quad \lambda = \frac{m}{n}, \;\;m,n \in {\mathbb N}, \tag{1} \] and \[ \dot x = -y + P(x,y), \quad \dot y = x + Q(x,y), \tag{2} \] in a neighborhood of the origin. Here, the natural numbers \(m, n\) are relatively prime, \(P, Q\) are polynomials, in the variables \(x,y \in {\mathbb R}\). A generalization of (1) and (2) is the planar complex system \[ \dot z = z + \sum_{k \geq 2} Z_k(z,w), \quad \dot w = -\lambda w - \sum_{k \geq 2} W_k(z,w), \quad \lambda = \frac{m}{n}, \;\;m,n \in {\mathbb N}, \tag{3} \] where \[ Z_k(z,w) = \!\! \sum\limits_{\alpha + \beta = k} a_{\alpha \beta} \, z^{\alpha} w^{\beta}\text{ and }W_k(z,w) = \!\! \sum\limits_{\alpha + \beta = k} b_{\alpha \beta} \, w^{\alpha} z^{\beta} \] are homogeneous polynomials with complex coefficients. If \(z, w \in {\mathbb R}\), and all coefficients \(a_{\alpha \beta}, b_{\alpha \beta} \in {\mathbb R}\), then the real system (3) coincides with (1). If \(\lambda = 1\) and the coefficients of (3) are complex conjugate: \(\overline{a_{\alpha \beta}} = b_{\alpha \beta}\), then (3) can be transformed into (2) by means of the transformation \(z = x + i y\), \(w = x - i y\), where \(i = \sqrt{-1}\). The main definitions are the following ones. System (3) is called \textit{normalizable} at the origin if there exists an analytic change of coordinates \(\xi = z + \dots\), \(\eta = w + \dots\), where dots are terms \(z^{\alpha} w^{\beta}\) of degrees \(\alpha + \beta \geq 2\), bringing the system (3) to the normal form \[ \dot \xi = \xi \biggl(1 + \sum_{k \geq 1} p_k U^k \biggr), \quad \dot \eta = -\lambda \eta \biggl(1 + \sum_{k \geq 1} q_k U^k \biggr), \quad U = \xi^m \eta^n. \tag{4} \] If the normal form (4) is linear, i.e., \(p_k = q_k = 0\) for all \(k \in {\mathbb N}\), then the initial system (3) is called linearizable. By definition, put \(\mu_0 = \tau_0 = 0\), and \(\mu_k = p_k - q_k\), \(\tau_k = p_k + q_k\) for all \(k \in {\mathbb N}\). Here, \(\mu_k\) is called the \(k\)th singular point quantity, and \(\tau_k\) is called the \(k\)-th generalized period constant. Clearly, system (3) is linearizable at the origin if and only if \(\tau_k = 0\) for all \(k \in {\mathbb N}\). The authors give a recursive algorithm for computing generalized period constants. It can be readily done using computer algebra systems (such as Mathematica or Maple). The authors study the linearizability problem for several systems (1) with quadratic polynomials \(P, Q\) and different resonances, i.e., different numbers \(m, n\). They obtain necessary and sufficient conditions of linearizability of such systems through the coefficients of \(P, Q\). Finally, they discuss linearizable conditions for several Lotka--Volterra systems, especially for the resonances \(\lambda = -\frac{3}{4}\) and \(\lambda = -\frac{3}{5}\).