Bifurcation of limit cycles in piecewise quadratic differential systems with an invariant straight line (Q2672983)

The study of piecewise differential systems enlarges one step further the difficulties of the study of ``simple'' ordinary differential equations. Even now, it is not still known the answer of the 16th Hilbert problem for a bound of the number of limit cycles in polynomial differential systems, except the trivial zero for the linear class. Some bounds as 4 for quadratic systems is in the minds of all researchers in this field, but no proof has already been given. When one enters into piecewise systems, then more parameters appear and even piecewise linear systems may already have limit cycles. The number of switching curves, their degree, and their relative position, altogether with the degree of the differential systems, will affect the number of limit cycles that may appear. In this paper, the authors study piecewise quadratic systems acting in two regions separated by a straight line. They put on each side a quadratic system of class II which is the second in a classification of 3 (I-III) of quadratic systems which may have limit cycles. Class III is the most generic, and class I the more degenerated. Class II thus stands in the middle. By locating the focus of both systems in the origin and on the splitting straight line, they try to combine the possible existence of limit cycles on one side with those from the other. Since by Hopf bifurcation, up to 3 limit cycles can be exploded from the focus, I guess that the authors had in mind the location of a case with 6 limit cycles. The great result is that they have obtained not 6 but even 7 limit cycles. Thinking a bit on retrospective, this is better understable if one thinks that piecewise linear systems may have limit cycles while ordinary linear systems cannot. So, the single piecewise property directly allows the creation of limit cycles. Anyway, it is very interesting the fact that they have proved the case with 7. And even more interesting to see that the number depends on the relative position of the splitting line with respect of the normal form used. In their case, the line being horitzontal leads up to 4 limits cycles, but being vertical allows up to 7. I wonder what will be the results if the more generic class III is ever used.