Will Random Cone-wise Linear Systems Be Stable?

TL;DR

This paper investigates the stability of random cone-wise linear systems near cusp-like equilibria, revealing non self-averaging Lyapunov exponents and cone trapping phenomena in high-dimensional stochastic dynamics.

Contribution

It introduces a model for multidimensional cone-wise linear dynamics with random matrices and analyzes stability properties, including finite-size effects and connections to diffusion persistence.

Findings

Lyapunov exponent is non self-averaging at large N

Systems can appear stable or unstable depending on initial conditions

Finite N effects cause cone trapping phenomena

Abstract

We consider a simple model for multidimensional cone-wise linear dynamics around cusp-like equilibria. We assume that the local linear evolution is either $\mathbf{v}^\prime=\mathbb{A}\mathbf{v}$ or $\mathbb{B}\mathbf{v}$ (with $\mathbb{A}$, $\mathbb{B}$ independently drawn a rotationally invariant ensemble of $N \times N$ matrices) depending on the sign of the first component of $\mathbf{v}$. We establish strong connections with the random diffusion persistence problem. When $N \to \infty$, we find that the Lyapounov exponent is non self-averaging, i.e. one can observe apparent stability and apparent instability for the same system, depending on time and initial conditions. Finite $N$ effects are also discussed, and lead to cone trapping phenomena.