Fixed-Time Input-to-State Stability (FxT-ISS)

• FxT-ISS is a property of nonlinear control systems defined by uniform, fixed-time convergence regardless of the initial state and external inputs.
• It leverages Lyapunov functions, small-gain theorems, and composite analysis to establish precise upper bounds on settling times across various system architectures.
• Applications include delay compensation in LTI systems, distributed control for parabolic PDEs, singularly perturbed dynamics, and real-time feedback optimization.

Fixed-Time @@@@1@@@@ (FxT-ISS) is a property of nonlinear control systems describing convergence of system trajectories to the origin, or to a bounded neighborhood determined by exogenous disturbances, within a constant time interval that is independent of initial conditions. FxT-ISS extends standard input-to-state stability (ISS) by strengthening the requirement on the settling time: while ISS and finite-time ISS only ensure boundedness or convergence in potentially state-dependent time, FxT-ISS guarantees uniform finite upper bounds on convergence time, making it crucial for control, optimization, and distributed systems requiring predictable, rapid performance even in the presence of persistent inputs or disturbances.

1. Formal Definitions and Lyapunov Characterizations

For a nonlinear system

$$\dot x = f(x,u), \quad x(0)=x_0, \quad u \in \mathcal{L}_\infty^m, \quad f(0,0)=0,$$

FxT-ISS is defined by the existence of comparison functions $\beta \in \KL$ and $\vartheta \in \K$ such that

$$|x(t)| \le \beta(|x_0|, t) + \vartheta(\|u\|_\infty), \quad \forall t \ge 0,$$

with the settling time $T_{\max} = \sup_{r \ge 0} T_\beta(r)$ uniformly bounded. The Lyapunov characterization requires a continuously differentiable $V: \R^n \to \R_{\ge0}$ satisfying:

• Sandwich bounds: $\ul\alpha(|x|) \le V(x) \le \ol\alpha(|x|)$ for class $\K_\infty$ functions $\ul\alpha,\ol\alpha$.
• Dissipation: $$V(x) \ge \chi(|u|) \implies \nabla V(x)^\top f(x,u) \le -\Psi(V(x))$$, where $\Psi(s) = a s^p + b s^q$ ($a, b > 0$, $p \in (0,1)$, $q > 1$).

A consequence is uniform settling time bounds: $T_{\max} \le \frac{1}{a(1-p)} + \frac{1}{b(q-1)},$ independent of initial condition magnitude (Tang et al., 28 Nov 2025, Tang et al., 24 Dec 2025, Tang et al., 2024).

2. Small-Gain Theorems for Interconnected Systems

The modern FxT-ISS literature provides Lyapunov-based small-gain theorems that generalize classic ISS small-gain results to fixed-time settings. In interconnections of $n$ subsystems, each subsystem is certified FxT-ISS via individual Lyapunov functions $V_i$ and comparison gains $\gamma_i$. The coupling is governed by nonlinear small-gain conditions: for strictly larger $\hat\gamma_i \in \K \cup \K_\infty$, the composition

$$\hat\gamma_1 \circ \hat\gamma_2(s) < s, \quad \forall s > 0$$

must hold.

Composite Lyapunov functions are constructed as

$$V(x) = \max\left\{ \sigma_\lambda(V_1(x_1)), \sigma_\lambda(\hat\gamma_1(V_2(x_2))), \sigma_\lambda(V_2(x_2)), \sigma_\lambda(\hat\gamma_2(V_1(x_1))) \right\},$$

with $\sigma_\lambda(s) = s^\lambda$, $\lambda \ge 1$ chosen to preserve dissipation. Dini-derivative calculus for the maximum yields

$D^+V(x) \le -\Psi(V(x)),$

thereby guaranteeing global uniform FxT-ISS for the entire interconnected network under suitable gain conditions (Tang et al., 28 Nov 2025, Tang et al., 24 Dec 2025).

3. FxT-ISS for LTI Systems and Delay Compensation

Research on fixed-time stabilization of linear time-invariant systems establishes feedback protocols for both delay-free and constant input-delay scenarios, enabling prescribed constant settling time for all nonzero initial conditions. The protocols often involve state-dependent homogeneous gains: $$u(t) = K_0 x(t) + K d(-\ln T) d(-\ln \|x(t)/\|x_0\|\|_d) x(t),$$ where $d(s)$ is a dilation associated with a homogeneity structure.

Quadratic Lyapunov functions parameterized by the prescribed settling time and initial state norm yield time invariance of the convergence window: $$\chi(t) = 1 - t/T, \quad x(t)=0 \text{ for } t \ge T.$$ Robustness to measurement noise and matched/additive disturbances is quantified via FxT-ISS estimates: $$\|x(t)\| \le \|x_0\| \beta(1, t-t_0) + \|x_0\| \gamma_1(\|q_1\|_{L^\infty} / \|x_0\|) + \|x_0\| \gamma_2(\|q_2\|_{L^\infty} / \|x_0\|).$$ For systems with input delays, Artstein's transform and predictor-based feedback preserve exact fixed-time convergence (Polyakov et al., 2023).

4. FxT-ISS in Distributed PDE Control

In the context of 1-D parabolic PDEs with destabilizing terms and boundary disturbances, FxT-ISS is achieved via boundary feedback controllers derived from the method of backstepping and system splitting. The stabilization strategy splits the state into disturbance-free and forced subsystems, applies invertible Volterra transformations, and designs controllers based on kernel equations ensuring exponential or prescribed-time decay for the disturbance-free part.

Lyapunov functionals quantify input-to-state effects, yielding bounds of the form: $$\|u[t]\|_{L^2} \le \beta(\|u_0\|_{L^2}, t) + \gamma(\|f\|_{L^\infty}) + \gamma_0(\|d_0\|_{L^\infty}) + \gamma_1(\|d_1\|_{L^\infty}),$$ with uniform fixed-time decay to zero in the absence of disturbances and boundedness in their presence (Zheng et al., 2024).

5. FxT-ISS in Singularly Perturbed Systems

Composite Lyapunov approaches extend FxT-ISS theory to multi-timescale singularly perturbed systems of the form: $$\dot x = f(x,z,u), \quad \varepsilon \dot z = g(x,z,u).$$ Two subsystems—the reduced slow dynamics and the boundary layer fast dynamics—are each certified FxT-ISS via their respective Lyapunov functions. Interconnection terms are bounded and must comply with quadratic-type inequalities and small-gain conditions: $$I_1 \le \nu_1 \tilde V_r^2 + \omega_1 \tilde V_b^2 + \rho_1(|u|), \quad \nu_1 < \frac{1}{2}\min k_{ir}.$$ Composite Lyapunov functionals

$$\Psi_\zeta(x,y) = \zeta V_r(x) + (1-\zeta) V_b(x,y)$$

permit the deduction of uniform fixed-time ISS for the full system, provided sufficient separation of timescales (sufficiently small $\varepsilon$) (Tang et al., 2024).

6. FxT-ISS in Feedback Optimization and Decentralized Games

FxT-ISS is foundational in real-time optimization and distributed control schemes where convergence speed and robustness to measurement errors or design perturbations are critical. Recent advances deploy nonsmooth, gradient-based controllers (homogeneous or composite) for minimization of time-varying cost functions with dynamic plants: $$\dot u = -\frac{F(x,u)}{|F(x,u)|^{\xi_1} - |F(x,u)|^{\xi_2}}, \quad F(x,u) = P^\top Q_1 x + Q_2 u + P^\top b_1 + b_2,$$ or for Nash-equilibrium seeking in potential games with FxT-ISS plant: $$\dot u_i = -\frac{G_i(x,u)}{|G_i(x,u)|^{\xi_1} + |G_i(x,u)|^{\xi_2}}.$$ Uniform fixed-time convergence is achieved without timescale separation, driven by quadratic Lyapunov analysis and verified by simulation. Finite-time bounds depend only on system and controller parameters and are independent of initial errors (Tang et al., 28 Nov 2025, Tang et al., 24 Dec 2025, Tang et al., 2024).

7. Common Themes, Methodology, and Practical Implications

The central technical theme in FxT-ISS research is the design and analysis of Lyapunov functions satisfying dual-power dissipation conditions and the construction of composite Lyapunov certificates in interconnected systems, PDEs, and multi-timescale dynamics. All approaches hinge on small-gain or cross-term inequalities, enabling uniform bounds on error and settling times.

Practical implications include:

• Predictable transient performance in feedback control and optimization.
• Robust distributed stabilization for systems with delays, disturbances, or time-varying parameters.
• Structural modularity in the analysis of networks, cascades, and games.

A unifying principle is that individual subsystems possessing FxT-ISS Lyapunov functions can be aggregated, via carefully tuned interconnection gains and composite functionals, to ensure robustness and settling time uniformity for the entire network or distributed control architecture. This approach is confirmed numerically across LTI systems, PDE boundary control, feedback optimization, and Nash game protocols.

---

Summary Table: FxT-ISS Principles and Applications

Area	Lyapunov Characterization	Core Results / Techniques
Interconnected ODE Systems	Sandwich bounds, dual-power decay	Small-gain theorem, composite Lyapunov, max-type
LTI & Delay Systems	Homogeneous, state-dependent gain	Prescribed time convergence, robustness estimates
Parabolic PDEs with Boundary Disturb.	Backstepping, splitting, kernels	ISS via energy, fixed-time via scheduling
Singularly Perturbed (Multi-rate)	Composite Lyapunov, small-gain	Uniform fixed-time under timescale separation
Feedback Optimization/Nash Games	Nonsmooth gradient/pseudogradient	Uniform fixed-time, spectral gap conditions

All results rely on rigorous construction and analysis of Lyapunov functions meeting fixed-time ISS criteria and appropriate nonlinear interconnection gain conditions (Tang et al., 28 Nov 2025, Tang et al., 24 Dec 2025, Tang et al., 2024, Polyakov et al., 2023, Zheng et al., 2024).