Stability Analysis of Nondifferentiable Systems

Abstract

Differential equations with right-hand side functions that are not everywhere differentiable are referred to as nondifferentiable systems. This paper introduces three novel methods to address stability issues in nondifferentiable systems. The first method extends the linearization method as it fails when the equilibrium is in a nondifferentiable region. We find that the stability of a piecewise differentiable system aligns with the behavior of its subsystems as long as the “distance” between these subsystems is sufficiently small. The second method is to examine the eigenvalues of the symmetric part of the Jacobian matrix in the vicinity of the equilibrium. This method applies to functions with even weaker regularity conditions, and does not require the eigenvalues to have a negative upper bound (or positive lower bound) over the domain. The third method establishes a connection between nondifferentiable systems and their approximate counterparts, revealing that their stability can be consistent under certain conditions. Additionally, we reaffirm the first two results via the approximation method. Examples are provided to illustrate the applications of our main results, including piecewise differentiable systems, general nondifferentiable systems, and realistic scenarios.