Recent advances about the uniqueness of the slowly oscillating periodic solutions of Wright's equation

From MaRDI portal
Publication:2269624



Abstract: An old conjecture in delay equations states that Wright's equation [ y'(t)= - alpha y(t-1) [ 1+y(t)], alpha in mathbb{R} ] has a unique slowly oscillating periodic solution (SOPS) for every parameter value alpha>pi/2. We reformulate this conjecture and we use a method called validated continuation to rigorously compute a global continuous branch of SOPS of Wright's equation. Using this method, we show that a part of this branch does not have any fold point nor does it undergo any secondary bifurcation, partially answering the new reformulated conjecture.


A Fourier expansion for periodic solutions of Wright's equation \[ \dot{x}(t)=-\alpha x(t-1)(1+x(t)) \] leads to a system of equations for the Fourier coefficients, with convolution terms. For \(\alpha\in[\frac{\pi}{2}+\epsilon, 2.3]\) \((\epsilon\approx 10^{-4})\), it is shown with computer assistance that this system can be solved uniquely (using the contraction principle) in a space of polynomially decaying sequences. Thus, a branch of periodic solutions parametrized with \(\alpha\) is obtained. The method is applied a second time to a homotopy between Wright's equation and the linear equation \[ \dot{x}(t)=-\frac{\pi}{2}x(t-1), \] to show that the constructed branch belongs to the continuum of the Hopf bifurcation at \(\alpha=\pi/2\).



Cites work


Cited in
(40)


Describes a project that uses

Uses Software






This page was built for publication: Recent advances about the uniqueness of the slowly oscillating periodic solutions of Wright's equation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2269624)