Construction of time-varying ISS-Lyapunov Functions for Impulsive Systems

This paper presents a method to construct time-varying ISS-Lyapunov functions from candidate ISS-Lyapunov functions for impulsive systems, thereby combining the ease of construction of the latter with the necessary and sufficient stability guarantees of the former, even in cases of simultaneous continuous and discrete instability.

The Gist

Imagine you are trying to keep a house safe from a storm. The "house" is a complex machine or system (like a robot, a power grid, or even a biological cell), and the "storm" represents outside disturbances or noise (like wind, power surges, or viruses).

In engineering, we want to know if the house will stay standing despite the storm. This concept is called Input-to-State Stability (ISS). If a system is ISS, it means: "No matter how hard the wind blows, the house might shake, but it won't collapse, and it will eventually settle back down."

To prove a house is stable, engineers usually use a "Safety Scorecard" called a Lyapunov Function. Think of this as a thermometer that measures how "unstable" the house is. If the temperature (instability) keeps going down, the house is safe.

The Problem: The "Candidate" Scorecard Has Blind Spots

For a long time, engineers used a specific type of scorecard called a Candidate ISS-Lyapunov Function. It's like a simple checklist:

1. The Flow: When the wind is blowing steadily, the house must get more stable.
2. The Jumps: When a sudden gust hits (an "impulse"), the house might shake violently, but it must settle down quickly enough before the next gust.

The Catch: This checklist only works if one part of the system is naturally stable.

• If the steady wind makes the house wobble (unstable flow) AND the sudden gusts make it wobble even more (unstable jumps), this old checklist gives up. It says, "I can't tell if this is safe," even if the house is actually fine. It's like a doctor saying, "I can't diagnose you because you have two different symptoms," when a cure actually exists.

The Solution: The "Time-Varying" Scorecard

The authors of this paper (Patrick Bachmann and Saeed Ahmed) introduced a new, super-powered scorecard called a Time-Varying ISS-Lyapunov Function.

The Analogy:
Imagine the old scorecard was a static map. It shows you the terrain, but it doesn't account for time. If you are walking through a swamp (unstable flow) and jumping over rocks (unstable jumps), a static map might tell you to stop because the terrain looks dangerous everywhere.

The new Time-Varying Scorecard is like a smart GPS with a timer.

• It knows that even if the terrain looks bad right now, if you keep moving for exactly 3 seconds, you will reach a safe spot.
• It accounts for the timing of the events. It knows that as long as the "jumps" happen at the right intervals, the system can survive even if both the flow and the jumps are individually chaotic.

This new tool is powerful because it provides a necessary and sufficient condition.

• Old way: "If the scorecard says yes, you are safe. If it says no, I don't know."
• New way: "If the scorecard says yes, you are safe. If you are safe, the scorecard will say yes." It never gives up on a system that is actually stable.

The Main Breakthrough: Building the New from the Old

The paper's biggest contribution is a construction method.

Usually, the new "Time-Varying" scorecards are very hard to build from scratch. They are like complex, custom-made suits of armor. The old "Candidate" scorecards are like standard, off-the-shelf t-shirts—easy to find and easy to make, but they have limitations.

The authors figured out how to sew the t-shirt into the armor.
They showed a mathematical recipe to take an easy-to-build "Candidate" scorecard (the t-shirt) and transform it into a powerful "Time-Varying" scorecard (the armor).

Why is this a big deal?

1. Ease of Use: Engineers can start with the simple, familiar tools they already know how to use.
2. Guaranteed Success: They can then upgrade those simple tools into the powerful new ones that are guaranteed to exist for any stable system, even the messy ones where everything is unstable at the same time.

Summary in Plain English

• The Goal: Prove that a system can handle chaos and noise without falling apart.
• The Old Tool: Good for simple cases, but fails when the system is chaotic in multiple ways at once.
• The New Tool: Works for everything, even the most chaotic systems, by looking at how things change over time.
• The Paper's Magic: It gives you a "DIY kit" to turn the simple, easy-to-make tools into the powerful, all-encompassing tools.

In short, the authors bridged the gap between "easy to build" and "guaranteed to work," allowing engineers to analyze and stabilize complex, chaotic systems that were previously considered too difficult to understand.

Technical Summary

Here is a detailed technical summary of the paper "Construction of time-varying ISS-Lyapunov Functions of Impulsive Systems" by Patrick Bachmann and Saeed Ahmed.

1. Problem Statement

The paper addresses the stability analysis of impulsive systems, a class of hybrid dynamical systems characterized by continuous evolution (flow) and abrupt state changes (jumps). The specific focus is on Input-to-State Stability (ISS) for these systems, particularly in infinite-dimensional Banach spaces.

Key Challenges Identified:

• Limitations of Candidate ISS-Lyapunov Functions: Existing methods using "candidate" ISS-Lyapunov functions (e.g., Dashkovskiy & Mironchenko, 2013) provide only sufficient conditions for ISS. They typically require that either the continuous flow or the discrete jumps be stable, with the other being unstable but bounded.
• Simultaneous Instability: Current candidate methods fail to provide conclusive stability results for systems where both the continuous dynamics (flow) and discrete dynamics (jumps) are simultaneously unstable.
• Lack of Converse Theorems: While time-varying ISS-Lyapunov functions were recently proposed (Bachmann et al., 2022) to provide necessary and sufficient conditions for ISS (even for simultaneously unstable systems), a practical method to construct these functions from existing, easier-to-find candidate functions was missing.

2. Methodology

The authors propose a constructive method to bridge the gap between candidate ISS-Lyapunov functions ($V_{cand}$) and time-varying ISS-Lyapunov functions ($V$).

Core Approach:

1. Definitions:
   • Candidate ISS-Lyapunov Function ($V_{cand}$): Satisfies standard Lyapunov conditions but allows for the derivative to be positive (unstable flow) or the jump map to increase the value (unstable jump), provided specific dwell-time or integral conditions are met.
   • Time-Varying ISS-Lyapunov Function ($V(t, x)$): A function that strictly decreases along trajectories outside a perturbation radius, regardless of whether the flow or jumps are individually unstable, provided the overall system is ISS.
2. Construction Strategy:
   The paper derives explicit formulas to transform a known candidate function $V_{cand}$ into a time-varying function $V(t, x)$. The construction involves:
   • Integrating the decay/growth rates defined in the candidate function conditions.
   • Introducing a time-dependent scaling factor that accounts for the interval between impulses ($t_{i+1} - t_i$).
   • Using a "max" operation or specific integral transformations to ensure the resulting function $V(t, x)$ satisfies the strict decrease condition required for time-varying ISS-Lyapunov functions.

Two Specific Scenarios Analyzed:

• Case A: Stable Flow, Unstable Jumps: The authors construct $V(t, x)$ using a combination of the candidate function and a time-decaying term derived from the integral of the decay rate $\rho$.
• Case B: Unstable Flow, Stable Jumps: A similar construction is provided where the time-varying function compensates for the unstable flow using the stability of the jumps.

3. Key Contributions

1. Construction Algorithm: The primary contribution is the explicit mathematical construction of a time-varying ISS-Lyapunov function $V(t, x)$ from a candidate ISS-Lyapunov function $V_{cand}(x)$. This allows researchers to leverage the simplicity of finding candidate functions while obtaining the rigorous guarantees of time-varying functions.
2. Necessary and Sufficient Conditions: By linking the two concepts, the paper reinforces that the existence of a time-varying ISS-Lyapunov function is a necessary and sufficient condition for ISS. This resolves the inconclusiveness of candidate functions for systems with simultaneous instability.
3. Extension to Infinite Dimensions: The results are formulated for systems in infinite-dimensional Banach spaces, covering a broader class of systems (e.g., PDEs with impulsive control) than many finite-dimensional counterparts.
4. Theoretical Unification: The paper unifies the "candidate" approach (easy to construct, sufficient only) with the "time-varying" approach (harder to construct, necessary and sufficient), showing that the latter can be systematically derived from the former.

4. Key Results

• Theorem 1 & 2 (Converse Theorem): The paper recalls and utilizes the result that an impulsive system is ISS if and only if a time-varying ISS-Lyapunov function exists.
• Theorem 4 (Stable Flow/Unstable Jumps): Provides the explicit formula for $V(t, x)$ when the flow is stable but jumps are unstable. The constructed function is defined as:
  $$V(t, x) := \max\{v_1(t, x), v_2(t, x)\}$$
  where $v_1$ involves an inverse integral transform of the candidate function adjusted by time, and $v_2$ prevents finite-time convergence issues.
• Theorem 6 (Unstable Flow/Stable Jumps): Provides the explicit formula for $V(t, x)$ when the flow is unstable but jumps are stable.
• Technical Lemmas (Appendix): The paper establishes rigorous technical properties, including:
  • Proof that ISS implies Weakly Uniformly Robustly Asymptotic Stability (WURS).
  • Lipschitz continuity of solutions with respect to initial conditions.
  • Existence of UGAS-Lyapunov functions for the associated feedback system.

5. Significance

• Overcoming Simultaneous Instability: The most significant impact is the ability to analyze and prove ISS for systems where both continuous and discrete dynamics are unstable. Traditional candidate methods cannot conclude stability in these cases, but the constructed time-varying functions can.
• Practical Utility: While time-varying Lyapunov functions are theoretically powerful, they are often difficult to guess or construct directly. This paper provides a "recipe" to generate them from standard candidate functions, making the powerful necessary-and-sufficient condition practically accessible.
• Standardization: The authors suggest that time-varying ISS-Lyapunov functions should become the standard formulation for impulsive system stability analysis, as their existence is guaranteed for any ISS system, whereas candidate functions may not exist or may be inconclusive.
• Robustness: The framework applies to infinite-dimensional systems, making it relevant for modern control applications involving distributed parameter systems (e.g., heat equations, wave equations) subject to impulsive disturbances or controls.

In summary, the paper solves a critical gap in hybrid systems theory by providing a constructive link between easily verifiable sufficient conditions (candidate functions) and rigorous necessary-and-sufficient conditions (time-varying functions), specifically enabling the stability analysis of impulsive systems with simultaneous continuous and discrete instabilities.