Periodic solutions near an equilibrium of a differential equation with a first integral (Q1106380): Difference between revisions

Let us consider a system of ordinary differential equations (1) \(\dot x=f(x)\) where f is a \(C^ 2\) map such that \(f(0)=0\), where it is assumed that \(Df(0)=A\) is an isomorphism and that (1) has a first integral \(G(x)=G_ 2(x,x)\) with \(G_ 2\) a nondegenerate bilinear form on \(R^ n\). If, moreover, \(\mu_ 1=i\beta\), \(\beta >0\) is a purely imaginary eigenvalue of A, and \(G_ 2\), restricted to the space \(E_{\mu}\), for some \(\mu_ k=k\mu_ 1\in \sigma (A)\), \(k\in N\), or to the space \(E_ 0=\oplus_{\mu \in \sigma (A),\mu =k\mu_ 1}E_{\mu}\), \(k\in N\) has a nonzero signature, the authors prove the existence of periodic solutions of (1) near the equilibrium point (\(\sigma\) (A) denote the set of all eigenvalues of A and \(E_{\mu}\) is, for each \(\mu\in \sigma (A)\), the generalized eigenspace of A corresponding to \(\mu)\). For it, they inject, in a standard way, the set of periodic solutions of (1) into the space \(R^ n\times [0,+\infty)\) and apply the approach of \textit{D. S. Schmidt} [J. Math. Anal. Appl. 63, 354-370 (1978; Zbl 0383.34026)] and \textit{J. Alexander} and \textit{J. Yorke} [Am. J. Math. 100, 263-292 (1978; Zbl 0386.34040)].