Fractional LaSalle Invariance Principle

• Fractional LaSalle Invariance Principle is a generalization of the classical invariance concept, designed for fractional dynamical systems with memory effects and uncertainty.
• It employs nonincreasing Lyapunov-type functionals and fractional derivatives, such as Caputo, to ensure that system trajectories converge asymptotically to equilibrium.
• The method leverages Mittag–Leffler decay and levelwise Hausdorff metrics in fuzzy systems to provide robust stability analysis even when energy decay is nonexponential.

The Fractional LaSalle Invariance Principle is a generalization of the classical LaSalle invariance principle to fractional-order dynamical systems, including systems with uncertainty such as fuzzy fractional differential equations. This extension incorporates the intrinsic memory effects and nonlocality associated with fractional derivatives—particularly of Caputo or Riemann–Liouville type—enabling rigorous asymptotic stability analysis in settings where the energy decay of trajectories is nonexponential and best described using Mittag–Leffler functions. The principle formulates conditions under which all solutions of such systems converge asymptotically to an equilibrium set, relying on nonincreasing Lyapunov-like functionals evaluated via fractional (often Caputo) differentiation, and, in the fuzzy case, interpreted through levelwise analysis of state cuts and Hausdorff metrics.

1. Mathematical Formulation and Statement

For a system governed by a Caputo fractional differential equation,

${}^cD_{0+}^q u(t) = f(u(t)),\qquad 0 < q < 1,$

with $u(t)$ taking values in a Banach space or, in the fuzzy case, in a suitable metric space $E$ of fuzzy numbers, the Fractional LaSalle Invariance Principle asserts: Let $V : E \to \mathbb{R}_+$ be a continuous Lyapunov-type functional such that along every solution,

${}^cD_{0+}^q V(u(t)) \leq 0.$

Define

$$\mathcal{S} = \{ u \in E : {}^cD_{0+}^q V(u) = 0 \}.$$

If every trajectory starting in $\mathcal{S}$ is bounded, and the largest invariant subset of $\mathcal{S}$ coincides with the equilibrium set $\mathcal{E}$, then every solution satisfies

$$\lim_{t \to \infty} \mathrm{dist}(u(t),\mathcal{E}) = 0.$$

Here, $\mathrm{dist}$ is typically Hausdorff distance when $u$ is a fuzzy state. This is an immediate extension of the classical LaSalle invariance theorem, now compatible with the nonlocal dynamics typical of fractional systems (En-naoui, 14 Sep 2025).

2. Key Assumptions and Analytical Framework

The application of the principle requires several key elements:

• Lyapunov Functionality: $V$ must be continuous with respect to the topology of the state space (e.g., levelwise in the fuzzy setting, or in a Banach space norm).
• Fractional Dissipativity: The fractional derivative (usually in the Caputo sense) ${}^cD_{0+}^q V(u(t))$ must be well-posed along trajectories and nonpositive.
• Invariant Set Characterization: The set $\mathcal{S}$ where fractional dissipation vanishes must contain no nontrivial invariant subsets besides the equilibrium set. This ensures there are no wandering or cyclic solutions within the zero-dissipation set.
• Boundedness: Solutions must remain bounded while in $\mathcal{S}$ to apply compactness arguments and to ensure accumulation points exist.

Lyapunov functionals are constructed to exploit system-specific properties, such as the nonexpanding nature of Mittag–Leffler decay rates in fractional systems or using Hausdorff distance as a “norm” in fuzzy-state spaces. This encompasses both deterministic and fuzzy fractional systems (En-naoui, 14 Sep 2025).

3. Proof Strategy and Structural Consequences

The proof mirrors the classical LaSalle scheme, modified for fractional and fuzzy settings. The principal steps are: a) Monotonicity: The functional $V(u(t))$ is nonincreasing due to ${}^cD_{0+}^q V(u(t)) \leq 0$, so the limit $V_\infty = \lim_{t\to\infty} V(u(t))$ exists and is finite. b) Compactness: Trajectories with bounded $V$ yield limit points in the completion of $E$ (e.g., completeness of the space with Hausdorff metric for fuzzy states). c) Limit Set Analysis: Any limit point $u^*$ must satisfy ${}^cD_{0+}^q V(u^*) = 0$ (verifiable, for instance, via integral argument for Caputo derivatives). d) Invariant Set Reduction: By the assumed invariance property, the only allowable limit set is the equilibrium set $\mathcal{E}$, forcing the trajectory to approach $\mathcal{E}$.

This yields convergence of all solutions to equilibrium, despite possible lack of strict monotonicity or exponential decay—critical for the analysis of systems exhibiting memory and/or uncertainty (En-naoui, 14 Sep 2025).

4. Extension to Fuzzy Fractional Systems

In fuzzy systems, the state $u(t)$ is a fuzzy number, and the Lyapunov functional $V$ must be compatible with the fuzzy metric structure, typically via levelwise Hausdorff distance. All relevant derivatives and integrals are taken on the $\alpha$-level representations, ensuring well-posedness for fuzzy-valued functions. The derivative condition, monotonicity property, and invariant set characterization hold “levelwise.” Analytical machinery such as the completeness of the metric space and continuity of $V$ with respect to the levelwise topology are fundamental in ensuring that every bounded trajectory produces limit points and that the principle applies (En-naoui, 14 Sep 2025).

5. Relevance to Stability and Mittag–Leffler Decay

Unlike classical systems where exponential stability is standard, the decay of solutions for fractional systems is most naturally described by the Mittag–Leffler function $E_q(-\lambda t^q)$. The invariance principle allows for stability proofs and asymptotic statements compatible with this slower, memory-laden decay, without requiring the existence of a strictly decreasing Lyapunov function. In the fuzzy case, the result unifies analysis under both uncertainty (through fuzzy sets) and nonlocality (through fractional derivatives), making it a critical tool for control and stability of advanced nonlinear or uncertain models (En-naoui, 14 Sep 2025).

6. Connections to Related Frameworks and Applications

The methods underlying the Fractional LaSalle Invariance Principle are consonant with a broad range of recent advances:

• Extension of Noether-type conservation laws for variational problems with fractional derivatives, where conserved quantities involving nonlocal integrals can serve as Lyapunov-type functionals in the stability analysis of fractional dynamical systems (Atanackovic et al., 2011).
• Invariant measure results for fractional stochastic models, wherein Lyapunov functionals adapted to dissipativity with respect to fractional noise identify invariant sets or measures, resonating with the Lyapunov–LaSalle paradigm for non-Markovian models (Gerencsér et al., 2020).
• Generalizations to infinite-dimensional and nonlocal gradient flows, where energy dissipation witnesses the asymptotic approach to equilibrium, and the set of stationary points is characterized via vanishing dissipation in the metric/topological space (Carrillo et al., 2020).

Fractional LaSalle principles, thus, serve as foundational results ensuring that the long-time behavior of systems with memory, uncertainty, nonlocality, or stochastic influences can be rigorously reduced to analysis on the invariant (often equilibrium) set, aiding in the design of robust control and stability guarantees for modern fractional systems.