Bifurcation analysis of a stochastically driven limit cycle (Q1729879)

The authors investigate a system with of a stochastically controlled limit cycle on a cylinder in the presence of white noise: \[ \begin{gathered} dy=-\alpha y dt+\sigma \sum_{i=1}^m (\vartheta)\circ dW_t ^i,\\ d\vartheta=(1+by)dt. \end{gathered}\tag{1} \] Here $(y,\vartheta) \in \mathbb{R} \times \mathbb{S}^1$ are cylindrical amplitude-phase coordinates, $ m\geq 2$ is a natural number, and $W_t ^i$ for $i \in {1,\dots, m}$ denote independent one-dimensional Brownian motions playing the role of noise of Stratonovich type. \par In the absence of noise $(\sigma = 0)$, the ordinary differential equation (1) has a globally attracting limit cycle at $y = 0$ if $\alpha 0$. In the presence of noise $(\sigma \neq 0)$, the amplitude is driven by phase-dependent noise. The real parameter $b$ induces shear: if $b \neq 0$, the phase velocity $\frac{d\vartheta}{dt}$ depends on the amplitude $y$. The stable limit cycle turns into a random attractor if $\sigma \neq 0$. The main problem of this paper concerns the study of the nature of this random attractor. The crucial quantity is the sign of the first Lyapunov exponent $\lambda_1 = \lambda_1(\alpha, b,\sigma)$ with respect to the invariant measure associated to the random attractor. \par The transition from negative to positive Lyapunov exponents for a fixed scattering parameter $\alpha$ and a sufficiently large noise $\sigma$ and/or a shift $b$ is established, which proves the existence of a bifurcation from random equilibrium to a random strange attractor. This results depend on a particular choice of the diffusion which makes an analytical treatment possible.\par Reviewer's remark: the results of this article should be extended to the problem of the stochastic Hopf bifurcation, where numerical studies indicate a shift from a negative to a positive Lyapunov exponent.