A stability-analysis method that constructs scalar Lyapunov functions—real-valued functions on the state space—whose sign and time derivative along system trajectories provide certificates of stability, asymptotic behavior, or instability for deterministic dynamical systems.

Principle

Search for a function V(x) that is positive definite (V(0)=0, V(x)>0 for x≠0) and has a nonpositive time derivative along solutions (dV/dt ≤ 0); negative definiteness of dV/dt yields asymptotic stability. The organizing idea is to convert trajectory statements into algebraic inequalities on V and its derivative.

Demonstration

Consider the linear ODE x' = Ax with equilibrium at 0. Choose V(x)=x^T P x with P symmetric positive definite. Then dV/dt = x^T (A^T P + P A) x. If there exists P>0 solving A^T P + P A = −Q with Q>0, then dV/dt = −x^T Q x < 0 for x≠0, proving asymptotic stability of 0.

Misapplication

Asserting global asymptotic stability from a Lyapunov function that is only locally positive definite, or using a candidate V that is not differentiable where trajectories pass, or treating Lyapunov functions as unique physical energies rather than certificates—each can lead to incorrect conclusions.

Consequence

When successfully applied, the method yields rigorous stability or asymptotic stability certificates, bounds on basins of attraction, invariance regions, and constructive controllers or estimators; it also supplies quantitative decay rates when dV/dt ≤ −cV.

Reversal

Instead of seeking a Lyapunov function that decreases, one can search for scalar witnesses of instability (a function that strictly increases along some trajectories) or inspect Lyapunov exponents and spectral properties that identify sensitivity rather than contractivity.

Boundary

Applies primarily to deterministic autonomous and time-varying ODEs, certain PDEs, and discrete-time maps with suitable modifications. It requires existence of appropriate V and sufficient regularity; it does not automatically handle noise-driven (stochastic) systems without extension, nor does it guarantee global results without global V.

Semantic Tension

Tension exists between viewing V as a mathematical certificate versus interpreting it as a physical energy; similarly, Lyapunov functions (finite-dimensional certificates) can be confused with Lyapunov exponents (trajectory growth rates) or with particular control-Lyapunov constructions that demand extra structure.

Synthesis

The Lyapunov Function Method is a constructive approach that reduces dynamical stability questions to finding scalar functions with sign and monotonicity properties; when such functions exist and satisfy derivative inequalities, they provide rigorous, often quantitative, conclusions about equilibrium stability and attraction.