On periodic solutions of even-order ordinary differential equations (Q1423484)

From MaRDI portal
scientific article
Language Label Description Also known as
English
On periodic solutions of even-order ordinary differential equations
scientific article

    Statements

    On periodic solutions of even-order ordinary differential equations (English)
    0 references
    4 March 2004
    0 references
    The authors investigate the problem of the existence and uniqueness of a periodic solution with a preassigned period \(\omega>0\) for the differential equation \[ u^{(2m)}=f(t,u,u',\ldots,u^{(m-1)}), \tag{1} \] where \(m\geq1\) and the function \(f:\mathbb{R}\times\mathbb{R}^{2m}\to\mathbb{R}\) satisfies the local Catathéodory conditions and is periodic with respect to the first argument with period \(\omega\). The results have the character of one-sided restrictions on \(f\) and comprise the case when \(f\) is a strongly increasing function of the phase variable, i.e., \[ \lim_{| x_1| +\cdots+| x_m| \to\infty} \frac{| f(t,x_1,\ldots,x_m)| }{| x_1| +\cdots+| x_m| }=+\infty \quad\text{for }t\in\mathbb{R}. \] From these results, the authors derive some sufficient conditions for the existence of a unique \(\omega\) periodic solution of the linear differential equation \[ u^{(2m)}=\sum_{k=1}^{m}p_{k}(t)u^{(k-1)}+p_{0}(t), \] which are different from those known so far. A typical result of the paper is given in the following theorem (\(L_{\omega}\) denotes the space of all \(\omega\)-periodic real functions which are integrable in the sense of Lebesgue on \([0,\omega]\).): Suppose that on \(\mathbb{R}\times\mathbb{R}^m\) the inequality \[ (-1)^{m-1}f(t,x_1,\ldots,x_m)\text{ \,sgn\,}x_1 \geq p(t)| x_1| ^{\lambda}- [p(t)]^\frac{1}{\lambda+1} \sum_{k=1}^{m-1}r_k| x_{k+1}| ^\frac{2\lambda}{\lambda+1} -q(t) \] holds, where \(\lambda>0\), \(r_k\geq0\), \(k=1,\ldots,m-1\), \(p\) and \(q\in L_{\omega}\) are nonnegative functions and \(p(t)\not\equiv0\). If, moreover, \[ \sum_{k=1}^{m-1}\left(\frac{\omega}{2\pi} \right)^\frac{2(m-k)\lambda}{\lambda+1} <(\lambda+1)\lambda^{-\frac{\lambda}{\lambda+1}}, \] then equation (1) has at least one \(\omega\)-periodic solution. Many results in the paper are in certain sense optimal.
    0 references
    periodic solutions
    0 references
    0 references
    0 references

    Identifiers