LaSalle's invariance principle

Consider an autonomous system ẋ = f(x) with a Lyapunov function V such that V(0) = 0, V(x) > 0 for x ≠ 0, and V̇(x) ≤ 0. The principle states that every bounded trajectory converges to the largest invariant set contained in the set {x : V̇(x) = 0}.² The proof uses the fact that V is nonincreasing along trajectories, so its limit exists; the omega-limit set of a bounded trajectory is nonempty, compact, invariant, and lies in the zero-derivative set, forcing convergence to that set.³

The principle is widely used in control theory and mechanical systems where constructing a negative definite derivative is difficult. For example, in the analysis of the pendulum with damping, the energy function has a derivative that vanishes at the upright equilibrium, yet the principle guarantees convergence to the stable equilibrium.⁴ Extensions include LaSalle's invariance principle for nonautonomous systems, where the conditions involve uniform continuity and the concept of a limiting set, and for discrete-time systems.⁵ The principle also underlies the Matrosov theorem, which relaxes conditions further by using auxiliary functions.

LaSalle's principle was inspired by earlier work of Barbashin and Krasovskii, who in 1952 proved a similar result for finite-dimensional systems, but LaSalle's formulation allowed for more general invariant sets and was independently developed.⁶ The principle is also valid in infinite-dimensional spaces, such as delay differential equations and partial differential equations, provided the trajectories are precompact.⁷ A subtle point is that the largest invariant set may contain multiple equilibria, and the principle does not specify which one is approached; additional analysis is needed to determine the limit. The principle has been applied to synchronization of coupled oscillators and to adaptive control, where the zero-derivative set often corresponds to a manifold of possible equilibria.

Joseph LaSalle, an American mathematician, introduced the principle in a 1960 paper in the Journal of Differential Equations, building on the work of Lyapunov and the Russian school.⁸ The principle became a cornerstone of nonlinear stability analysis, bridging the gap between Lyapunov's theory and practical engineering applications. It is taught in standard textbooks on nonlinear systems and is a key tool in the design of observers and controllers. The principle's elegance lies in its simplicity: it replaces the stringent requirement of negative definiteness with a geometric condition on the trajectory's limit set, making it indispensable for analyzing systems with conserved quantities or symmetries.

Lyapunov function

A scalar function that is positive definite and has a nonpositive derivative along trajectories, used to certify stability.

Invariant set

A set of states such that any trajectory starting in the set remains in it for all time.

Omega-limit set

The set of all accumulation points of a trajectory as time tends to infinity.