• PDF
• Cite
• Share
Article Contents  Article Contents

# Generalised Lyapunov-Razumikhin techniques for scalar state-dependent delay differential equations

• * Corresponding author: A. R. Humphries
• We present generalised Lyapunov-Razumikhin techniques for establishing global asymptotic stability of steady-state solutions of scalar delay differential equations. When global asymptotic stability cannot be established, the technique can be used to derive bounds on the persistent dynamics. The method is applicable to constant and variable delay problems, and we illustrate the method by applying it to the state-dependent delay differential equation known as the sawtooth equation, to find parameter regions for which the steady-state solution is globally asymptotically stable. We also establish bounds on the periodic orbits that arise when the steady-state is unstable. This technique can be readily extended to apply to other scalar delay differential equations with negative feedback.

Mathematics Subject Classification: Primary: 34K20, 37L45.

 Citation: • • Figure 3.  (a) The green region is the parameter region where the zero solution was proven to be locally asymptotically stable using Lyapunov-Razumikhin techniques in  (denoted as the set $\{P(1,0,3)<1\}$). The blue curve shows the boundary of the region where we numerically evaluate $M_{1,\infty} = N_{1,\infty} = 0$ and the red curve shows the boundary of the set where $M_{2,\infty} = N_{2,\infty} = 0$. The steady state of (2) is guaranteed to be globally asymptotically stable to the left of the red curve. The brown curve shows the locus of a fold bifurcation of periodic orbits computed using DDE-Biftool. To the right of the brown curve there are co-existing non-trivial periodic orbits and the steady-state cannot be globally asymptotically stable. Other parameters are fixed at $a = c = 1$. (b) Bifurcation diagram at $\mu = -0.25$ as $\sigma$ is varied showing a subcritical Hopf bifurcation from the steady state and the amplitude of the resulting periodic orbits. There is a small interval of $\sigma$ values for which bistability occurs, and the steady-state is only locally asymptotically stable.

Figure 4.  (a) & (c): Numerically simulated solutions $u(t)$ to (2) with $\varphi (t) = 0.99N_0$ for all $t\leq0$, along with the bounds $M_{k,n}$ and $N_{k,n}$ from Theorem 3.6 with $k = 1$ and $k = 2$. In (a), $(\mu,\sigma) = (-0.25,-1.75)\not\in \mathop \Sigma_{w}$ and the steady state solution is unstable. The solution $u(t)$ converges to a periodic orbit. In (c), $(\mu,\sigma) = (-2,-2.8)\in \mathop \Sigma_{w}$ and $u(t)$ converges to the steady state. (b) & (d): For the same values of $\mu$ as in (a) and (c) we vary $\sigma$ and plot the bounds $[M_{k,\infty},N_{k,\infty}]$ on the persistent dynamics from Theorem 3.7 for $k = 1$ and $k = 2$ along with $[\min_t u(t),\max_t u(t)]$ for the periodic orbit created at the Hopf bifurcation when $\sigma$ crosses the boundary of $\mathop \Sigma_{w}$.

Figure 5.  Relative improvement in the lower and upper bounds of periodic solutions to (2) using the $k = 1$ and $k = 2$ generalised Lyapunov-Razumikhin technique. The yellow region matches the region of global asymptotic stability indicated by the blue (for $k = 1$) and red (for $k = 2$) curves in Figure 3

Figure 1.  (a) The analytic stability region $\Sigma_\star$ in the $(\mu,\sigma)$ plane, divided into the delay-independent cone $\mathop \Sigma _{\Delta}$, and the delay-dependent wedge $\mathop \Sigma_{w}$ and cusp $\mathop \Sigma_{c}$. (b) Sample dynamics using parameter values in the $\mathop \Sigma _{\Delta}$ (upper panel) and $\mathop \Sigma_{w}$ (lower panel). Both state-dependent ($c = 1$) and constant delay ($c = 0$) solutions are shown in each case. The initial function is the constant function $\varphi(t) = 1$ for all the examples.

Figure 2.  Lower bounds on $\delta$, with $a = c = 1$ for parameter values $(\mu,\sigma)$ in parts of the wedge $\mathop \Sigma_{w}$ so that the ball $\{\varphi :\|\varphi \|<\delta\}$ is contained in the basin of attraction of the steady state. These bounds were derived using the Lyapunov-Razumikhin techniques presented in . The SDDE is initially integrated through $k\tau$ time units so that bounds can be established on derivatives of the solution. The value of $k\in\{1,2,3\}$ which results in the largest bound is indicated

•  Open Access Under a Creative Commons license

## Article Metrics  DownLoad:  Full-Size Img  PowerPoint