Divergence of a real normalizing transformation

From MaRDI portal





The following system is considered: (1) \(\dot z{}_ j=z_ j(\lambda_ j+f_ j(z_ 1,z_ 2))\), \(j=1,2\), where \(f_ j\) are analytic at zero \(z_ 1=0\), \(z_ 2=0\), \(f_ j(0,0)=0\) and \(\lambda_ 1=-\lambda_ 2\neq 0\). There exists a formal coordinate transformation (2) \(z_ j=w_ j(1+h_ j(w_ 1,w_ 2)), j=1,2\), reducing the system (1) to the normal form (3) \(\dot w_ j=w_ j(\lambda_ j+g_ j(w_ 1w_ 2)), j=1,2\), where \(h_ j\) and \(g_ j\) are formal power series of \(w_ 1,w_ 2\) without free members and \(g_ j\) depends only on the product \(w_ 1w_ 2\). Here only the case Re \(\lambda {}_ 1=Re \lambda_ 2\) is considered. The following theorem is proved: For each analytic normal form (3) satisfying some reality conditions, there exists an analytic system (1), formally equivalent to system (3) and for which any normalizing transformation (2) is divergent.











This page was built for publication: Divergence of a real normalizing transformation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q796711)