On periodic solutions of certain \(n\)th order differential equations (Q1910083)

From MaRDI portal
scientific article
Language Label Description Also known as
English
On periodic solutions of certain \(n\)th order differential equations
scientific article

    Statements

    On periodic solutions of certain \(n\)th order differential equations (English)
    0 references
    0 references
    0 references
    29 August 1996
    0 references
    Extensive periodicity theory exists for second other differential equations and for higher order equations that fit the form \(x' = Ax\), where \(A\) is a matrix, such as the Floquet theory. For second order systems written as \(x' = f(x,y)\), \(y' = g(x,y)\) the Poincaré-Dulac, Bendixson-Dulac theorems, which may be regarded as fixed point theorems, or the phase portrait approach, appear to be the classic techniques used in search of periodicity or almost periodicity, as defined by Besicovich. For the more general equations of the type \(x' = f(x, \varepsilon)\) the Poincaré method of small parameters has a devoted popular following. Some of the more sophisticated topological arguments based on the use of Lyapunov functions in a compact subdomain, have been combined with the Brouwer, Lefschetz or Schauder fixed point theorems. The authors are professors at the Jagiellonian University of Kraków, the Alma Mater of the professors Stanisław Zaremba and Tadeusz Wa. zewski, originator of much theory based on topological arguments, particularly making use of Borsuk's theory of retracts. The purpose of the paper reviewed here is to derive an alternate technique mainly based on the Lefschetz fixed point theorem and the generally neglected, but very useful Ważewski's retract principle. The argument proceeds along the following lines: Theorem 0: Consider the Cauchy problem \(x' = f(t,x)\), where \(f : \mathbb{R}^n \times \mathbb{R} \to \mathbb{R}\) is continuous and \(T\)-periodic. Let \(E\) and \(E^{\text{exit}}\) be a pair of compact Euclidean neighborhood retracts in \(\mathbb{R}^n\). Let \(t \to \Phi_{(t_0, t - t_0)}\) represent the solution map. \((\Phi_{(0,t)}\) is also called Poincaré's operator.) Assume the existence of a sequence of constants \(\varepsilon_k > 0\), such that \(\lim \varepsilon_k = 0\), as \(k \to \infty\). Also assume that \(\mathbb{R} \times E^{\text{exit}}\) is contained in \(\mathbb{R} \times E\), and \(\Phi_{(t, \varepsilon k)} \notin E\). Also \(\kappa (E) - \kappa (E^{\text{exit}})\) is not equal to zero, where \(\kappa\) is the Euler characteristic. Then there exists a \(T\)-periodic solution staying in the set \(E\), and the set of fixed points \(K_E\) in \(E\) is compact and isolated in the set of fixed points of the Poincaré operator, and the fixed point index: \(\text{ind} (\Phi_{(0,T)}, K_E) = \kappa (E) - \kappa (E^{\text{exit}})\). Assuming that a solution exists to the Cauchy problem stated above, the authors prove two theorems concerning the equation \(y^{(n)} + a_{n - 1} y^{(n - 1)} + \cdots + a_1 y' = b(t,y,y', \dots, y^{(n - 1)})\), where \(a_i\) are real numbers, and \(b\) is a continuous function. One is a corollary to a theorem of \textit{J. R. Ward jun.} [Proc. Am. Math. Soc. 81, 415-420 (1981; Zbl 0461.34029)], but proved by an entirely different technique. The second one with conditions identical to Theorem 1, (except for the assumption of existence of the \(T\)-periodic solution), proves that in fact at least two periodic solutions exist. The theorem is new. The technique of proof needs to be studied by people working on similar problems.
    0 references
    0 references
    periodic solution
    0 references
    Lefschetz fixed point theorem
    0 references
    Ważewski's retract principle
    0 references

    Identifiers