Uniqueness of algebraic limit cycles for quadratic systems (Q5946932)

All known quadratic systems (QS) having an algebraic limit cycle are contained in five families. The degree of the algebraic curve containing the limit cycle is 4 in four of these families and 2 for the other one. Furthermore, it is also known that if there is another QS having an algebraic limit cycle, it should have at least degree 5. The main result in this paper is that for all these five families of QS the only limit cycle is the algebraic one. This result was known just for one case, the one in which the degree of the algebraic limit cycles is 2. In order to give a brief idea of its proof, recall that an irreducible algebraic curve \(f=0\) is called invariant algebraic curve for a QS, \(X=(p,q),\) if there exists a one-degree polynomial \(k\), called the cofactor of \(f,\) such that \(\dot f=f_xp+f_yq=kf.\) For two of the studied families the proof of the stated result follows by finding a kind of Lyapunov function of the form \(H(x,y)=r(x,y)|f(x,y)|^{1/2},\) with \(r(x,y)\) a polynomial and such that \(\dot H\) does not change sign on each connected component of \({\mathbb{R}}^2\setminus\{f=0\}.\) The proof in the other three cases is a corollary of the following nice result, also proved in the paper: Let \(f(x,y)\) be a real invariant algebraic curve, with degree larger than one and cofactor \(k,\) for the QS, \(X=(p,q),\) being \(p\) and \(q\) coprimes. Let \(P(x,y,z)=z^2p(x/z, y/z)\) (respectively \(Q(x,y,z), K(x,y,z)\)) be the homogenization of \(p(x,y)\) (respectively \(q(x,y), k(x,y)\)). Suppose that there are at two common zeros in \({\mathbb C}P^2\) of \(P, Q\) and \(K.\) Then all limit cycles of the QS are contained in the algebraic curve \(f=0.\)