On the Stability of Discrete Reaction-Diffusion System of Networked Dynamical Systems
Dinesh Kumar
TL;DR
The paper addresses the stability of discrete reaction-diffusion networks of $m$ coupled dynamical systems in $n$ variables around a homogeneous equilibrium, formalized as $\dot{x}=f(x)-Lx$ with $L=\mathcal{L}_1\oplus\dots\oplus\mathcal{L}_n$. The approach linearizes around the equilibrium to $\dot{x}=(Df(\bar{x})-L)x$, then uses a block-diagonalization via $x=Py$ to obtain $\dot{y}=(P^{-1}Df(\bar{x})P-\Lambda)y$ and applies Gershgorin disc arguments together with Weyl's monotonicity. The main result provides a sufficient condition: the averaged local Jacobians must be diagonally dominant with negative diagonal entries, and the network connectivity must satisfy $\lambda_2\ge\tau$. This is illustrated with ecological metapopulation networks, showing that stronger connectivity can stabilize homogeneous equilibria even when local dynamics are unstable, offering a design principle for robust stability in spatially structured systems.
Abstract
Paper provides a sufficient condition for the local stability of the reaction-diffusion system of networked dynamical systems at its homogeneous equilibrium point. It is illustrated through an ecological meta-populations (spatial-structured populations) network of predator-prey systems.
