🔢 mathematics

Unified Lyapunov Method for ISS of PDEs: A Tutorial on Constructing Generalized Lyapunov Functionals for Parabolic and Hyperbolic Equations

This tutorial presents a unified generalized Lyapunov method for establishing input-to-state stability in $L^q$ spaces by systematically constructing input-dependent functionals for nonlinear parabolic equations, first-order nonlinear hyperbolic equations, and second-order linear wave equations with boundary disturbances.

Original authors: Jun Zheng, Guchuan Zhu

Published 2026-05-05

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jun Zheng, Guchuan Zhu

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to keep a complex machine running smoothly, like a giant, invisible drum or a long, vibrating rope. This machine is constantly being pushed and pulled by outside forces—wind, random bumps, or someone shaking the ends. In engineering and math, we call these outside forces "disturbances."

The big question is: If you push the machine, how much will it wobble? And will it eventually settle down, or will it go crazy?

This paper is a "how-to guide" for answering that question for a specific type of machine described by Partial Differential Equations (PDEs). These are fancy math equations used to model things like heat spreading through a metal rod, traffic flow, or sound waves.

Here is the breakdown of the paper's main ideas, using simple analogies.

1. The Problem: The Old Rules Don't Work for "Pushy" Boundaries

For a long time, engineers used a tool called the Classical Lyapunov Method (think of this as a "Stability Scorecard").

How it worked: You pick a number (the score) that represents how much energy the system has. If the score goes down over time, the system is stable.
The Limitation: This worked great when the "pushes" happened inside the machine (like heat generated inside a room). But, it fell apart when the pushes happened at the edges (like someone shaking the door of the room).
The Analogy: Imagine trying to measure the stability of a boat. If the waves hit the boat from the inside (a leak), the old scorecard works fine. But if someone is standing on the dock and violently shaking the boat by the railing (a boundary disturbance), the old scorecard gets confused. It can't account for the force coming from the outside edge, so it can't tell you if the boat will capsize.

2. The Solution: The "Generalized" Scorecard

The authors introduce a new tool called the Generalized Lyapunov Method (GLM).

The Innovation: Instead of just looking at the boat's current energy, the new scorecard explicitly includes the force of the person shaking the railing in its calculation.
The Metaphor: Think of the old method as a thermometer that only measures the temperature of the water. The new method is a thermometer that also measures how hard the wind is blowing. By combining these two, you get a much more accurate picture of what's happening.
The Result: This new method allows mathematicians to prove that even if someone is shaking the edges of the system, the system will stay under control (mathematically called Input-to-State Stability, or ISS).

3. The "Secret Sauce": Truncation (The "Cut-Off" Trick)

How did they build this new scorecard? They used a clever mathematical trick called Stampacchia's truncation.

The Analogy: Imagine you are trying to balance a stack of books, but a giant is occasionally dropping heavy anvils on top. Instead of trying to calculate the exact weight of every single anvil (which is impossible), you decide to say, "Any weight over 100 pounds is just '100 pounds' for our calculation."
In the Paper: They create a "cutoff" value ( $M$ ). If the disturbance (the shaking) is stronger than this cutoff, the math treats it as a constant, manageable limit. This allows them to ignore the messy, infinite details of the disturbance and focus on the core stability of the system.

4. What They Tested (The Three Examples)

To prove their new method works, they applied it to three very different types of "machines" (equations):

The Heat Equation (Parabolic): Imagine heat spreading through a 3D block of metal.
- The Challenge: The edges of the metal are being heated or cooled by random external sources.
- The Result: They showed that no matter how the edges are heated, the temperature inside stays predictable and bounded.
The Traffic Flow Equation (First-Order Hyperbolic): Imagine cars moving on a highway where the speed depends on how many cars are on the road.
- The Challenge: Cars are entering or leaving the highway at a rate that changes randomly.
- The Result: They proved the traffic flow won't jam up infinitely; it will settle into a manageable pattern.
The Wave Equation (Second-Order Hyperbolic): Imagine a guitar string or a drumhead.
- The Challenge: The ends of the string are being pulled and pushed, and there is friction (damping) trying to stop the vibration.
- The Result: They showed that even with the ends being wiggled, the string's vibration will die down over time, provided the friction is strong enough.

5. The Bottom Line

The paper doesn't just say "this works." It provides a step-by-step recipe for building these new "Generalized Scorecards" for different types of equations.

Why it matters: Before this, if you had a complex system with messy edges, you might have had to use very difficult, specialized math tools that only worked for one specific shape or material.
The Takeaway: The authors have created a unified toolkit. Whether you are dealing with heat, traffic, or sound waves, if you have disturbances at the boundaries, you can now use this specific "Generalized" method to prove the system is safe and stable.

In short: They fixed a broken ruler (the old method) so it can measure systems that are being pushed from the outside, and they showed you exactly how to use the new ruler on three different types of problems.

Technical Summary: A Unified Lyapunov Method for ISS of PDEs

Problem Statement

The paper addresses the challenge of establishing Input-to-State Stability (ISS) for Partial Differential Equations (PDEs) subject to external disturbances, with a specific focus on boundary disturbances. While ISS theory is well-developed for finite-dimensional systems and PDEs with distributed (in-domain) disturbances, the analysis of PDEs with boundary disturbances remains difficult. This difficulty arises because boundary disturbances introduce unbounded input operators in the abstract formulation of the system, creating significant obstacles for standard stability analysis.

Existing methods for handling boundary disturbances include eigenfunction expansions, semigroup theory, monotonicity methods, and the De Giorgi iteration. However, these approaches often suffer from limitations:

Linear operator-based methods (e.g., eigenfunction expansions) are typically restricted to one-dimensional linear PDEs.
Classical Lyapunov methods (CLM) based on Sobolev embeddings struggle with Dirichlet-type boundary disturbances, as the resulting boundary terms in the Lyapunov derivative cannot be easily handled.
Many existing techniques are not readily extendable to nonlinear PDEs or higher-dimensional domains.

Consequently, there is a need for a unified framework capable of handling diverse PDE classes (parabolic, hyperbolic, linear, nonlinear) and various boundary conditions (including Dirichlet) without relying on complex spectral analysis or admissibility conditions.

Methodology: The Generalized Lyapunov Method (GLM)

The paper proposes and demonstrates the Generalized Lyapunov Method (GLM), which relies on the construction of Generalized ISS-Lyapunov Functionals (GISS-LFs).

Core Concept

Unlike classical ISS-Lyapunov functionals (CISS-LFs), which depend solely on the system state $x$ , a GISS-LF, denoted as $\bar{V}(x, u)$ , explicitly incorporates the external input (disturbance) $u$ . The functional is constructed as a sum of classical Lyapunov candidates applied to transformed states:
$\bar{V}(x, u) = \sum_{i \in I} V(h_i(x, u))$
where $h_i(x, u)$ are continuous functions that shift the state based on the magnitude of the disturbance.

Key Technical Steps

Truncation Functions: The method utilizes Stampacchia's truncation technique. For a given disturbance bound, a truncation function $g(s)$ and its integral $G(s)$ are defined. These functions are zero for negative arguments and behave like power functions ( $s^p$ ) for positive arguments.
Disturbance-Dependent Shifts: A constant $M$ is constructed based on the supremum norms of the disturbances (e.g., $M = \phi^{-1}(\|d\|_\infty) + \dots$ ). The GISS-LF is then formed by evaluating the Lyapunov candidate on shifted states, such as $V(w - M)$ and $V(-w - M)$ .
Derivative Analysis: The time derivative of the GISS-LF is computed using integration by parts (or Green's formula for higher dimensions). The specific structure of the truncation functions and the choice of $M$ ensure that boundary terms involving disturbances become non-positive or cancel out.
Inequalities: The analysis relies on fundamental differential inequalities (Agmon, Poincaré, Jensen, Young, Gronwall) and specific properties of the truncation functions (properties G1–G8) to derive decay rates and ISS estimates.

Key Contributions and Results

The paper provides a systematic tutorial on constructing GISS-LFs for three representative classes of PDEs, establishing ISS estimates in $L^q$ spaces for any $q \in [2, \infty]$ .

1. N-Dimensional Nonlinear Parabolic Equations

System: A higher-dimensional parabolic equation with mixed nonlinear boundary disturbances (Dirichlet on $\Gamma_1$ , Robin/Neumann on $\Gamma_2$ ) and distributed in-domain disturbances.
Construction: The GISS-LF is constructed as $\bar{V}(w) = V(w - M) + V(-w - M)$ , where $V(w) = \int_\Omega G(w) dy$ .
Result: The system is proven to be ISS in $L^q$ norms. The estimate shows exponential decay of the initial state influence and a bounded gain from the disturbances.

2. 1-D First-Order Nonlinear Hyperbolic Equations

System: A hyperbolic equation modeling semiconductor manufacturing, where the velocity depends on the total mass (integral of the state).
Construction: A weighted GISS-LF is used: $\bar{V}(\rho) = V(\rho - M) + V(-\rho - M)$ , with an exponential weight $e^{-ry}$ to handle the boundary coupling.
Result:
- Global ISS: Established under a constant lower bound assumption on the velocity function.
- Local ISS: Established for general velocity functions, where the decay rate depends on the bound of the initial state and disturbance.
- Estimates are derived for both local and global $L^q$ norms.

3. 1-D Second-Order Linear Hyperbolic Equations (Wave Equation)

System: A wave equation with boundary velocity damping and boundary disturbances.
Construction: The state is decomposed into characteristic variables ( $w_t \pm c w_y$ ). The GISS-LF is a sum of four terms involving these characteristic variables shifted by $M$ :
$\bar{V}(w) = \sum V_i(\text{characteristic variable} \pm M)$
Result: The method successfully handles the boundary damping and disturbance terms simultaneously. Explicit ISS estimates are derived in $L^q$ norms. Notably, the paper highlights that the disturbance-free system is superstable (finite-time stable) under this framework.

Significance and Claims

The paper positions the GLM as a significant advancement over the Classical Lyapunov Method (CLM) for PDEs with boundary disturbances.

Handling Dirichlet Disturbances: The primary claim is that GLM overcomes the specific limitation of CLM regarding Dirichlet boundary disturbances. By explicitly incorporating the disturbance bound into the Lyapunov functional via the shift $M$ , the method avoids the "obstacles" encountered by classical approaches where boundary terms cannot be bounded.
Unified Framework: The authors claim that GLM offers a unified approach applicable to parabolic, first-order hyperbolic, and second-order hyperbolic equations, covering both linear and nonlinear cases.
Simplicity and Efficiency: The method relies primarily on integration by parts and standard inequalities, circumventing the need for:
- Verifying admissibility conditions for input operators (required in semigroup approaches).
- Complex spectral analysis or eigenfunction expansions (required in Riesz-spectral approaches).
Flexibility: The approach allows for the construction of ISS estimates in arbitrary $L^q$ spaces ( $q \in [2, \infty]$ ), providing a more comprehensive characterization of system robustness than methods restricted to $L^2$ .

The paper concludes that while constructing a universally applicable GISS-LF for all PDEs remains an open challenge due to the diversity of PDE structures, the GLM provides a powerful and systematic tool for the most representative classes of PDEs, paving the way for future applications in controller design and observer synthesis.