Parameter functionalization method in an approximate calculation of weak self-oscillation modes (Q917757)

The author considers the differential equation \(\dot x=A(\lambda)x+\phi (x;\lambda)\) where \(x\in {\mathbb{R}}^ N\), A(\(\lambda\)) is a constant (N\(\times N)\)-matrix continuously depending on a real parameter \(\lambda\), \(| \lambda -\lambda_ 0| \leq a_ 0\), the vector function \(\phi\) (x;\(\lambda\)) is continuous in both of its variables and satisfies the inequality \(\| \phi (x;\lambda)\| \leq \| x\| \delta (\| x\|),\) \((x\in {\mathbb{R}}^ N,| \lambda -\lambda_ 0| \leq a_ 0)\), and where \(\delta\) (\(\epsilon\)) is a continuous, monotonic non-negative function defined for \(\epsilon\geq 0\) and satisfying \(\delta\) (\(\epsilon\))\(\to 0\) for \(\epsilon\to 0\). If the matrix \(A(\lambda_ 0)\) has two simple complex-conjugate eigenvalues \(\pm 2\pi i\) and if the numbers \(0,\pm 2k\pi i\) \((k=2,3,...)\) are not its eigenvalues then the above equation has a periodic solution for small \(\lambda -\lambda_ 0\). For any positive \(\epsilon\) there exist numbers \(\lambda_{\epsilon}\), \(| \lambda_{\epsilon}-\lambda_ 0| <\epsilon\), and \(T_{\epsilon}\), \(| T_{\epsilon}-1| <\epsilon\), such that for \(\lambda =\lambda_{\epsilon}\) the above system has a \(T_{\epsilon}\)-periodic solution \(x_{\epsilon}(t):\| x_{\epsilon}(t)\| <\epsilon\) for \(-\infty <t<\infty\). The author presents an iterative procedure approximating the solution \(x_{\epsilon}(t)\) and the values of the period \(T_{\epsilon}\) and the parameter \(\lambda_{\epsilon}\).