🔢 mathematics

Quantitative rapid boundary stabilization via modal decomposition and its application to the Allen-Cahn equation

This paper establishes quantitative rapid boundary stabilization for the one-dimensional Allen-Cahn equation by developing a modal decomposition approach that explicitly links feedback laws and stabilization costs to the prescribed decay rate, while also deriving null controllability and finite-time stabilization results.

Original authors: Shengquan Xiang, Yu Xiao, Can Zhang

Published 2026-07-07

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Shengquan Xiang, Yu Xiao, Can Zhang

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 have a long, thin, flexible rope (representing a physical system like heat flowing through a metal rod or a chemical spreading through a liquid). This rope is naturally unstable; if you nudge it, it might start to wiggle wildly and never settle down. Your goal is to hold one end of the rope and apply just the right amount of force to stop the wiggling as quickly as possible, no matter how hard it was shaking to begin with.

This paper is about a new, highly precise way to figure out exactly how much force to apply and how fast the rope will stop moving.

Here is the breakdown of their discovery using simple analogies:

1. The Problem: The "Unruly Rope"

The authors are studying a specific type of "rope" called the Allen–Cahn equation. Think of this as a rope that doesn't just wiggle randomly; it has a habit of getting stuck in certain shapes or flipping between states (like a switch that gets stuck in the "on" or "off" position).

In control theory, we want to apply a "feedback law." This is like a smart hand at the end of the rope that feels the movement and pushes back instantly to stop it. The goal is rapid stabilization: making the rope stop moving at a speed we choose (let's call this speed $\lambda$ ). The faster we want it to stop, the harder the hand has to push.

2. The Old Way vs. The New Way

The Old Way (Qualitative):
Previously, scientists knew they could build a hand that stops the rope quickly. They used a method called Modal Decomposition. Imagine the rope's movement is a song made of many different musical notes (frequencies).

High notes (High Frequencies): These wiggles die out naturally very fast, like a high-pitched squeak that fades instantly. You don't need to do much here.
Low notes (Low Frequencies): These are the deep, slow, stubborn wiggles that keep the rope moving. You need to actively stop these.

Old methods could prove you could stop the low notes, but they were vague about how much force was needed. It was like saying, "You can stop the car, but I can't tell you how much gas you'll need to hit the brakes."

The New Way (Quantitative):
This paper does something new: it calculates the exact cost. It answers the question: "If I want the rope to stop 10 times faster, how much more force do I need?"

They found that the force required grows in a very specific, predictable way (mathematically, it looks like $e^{\sqrt{\lambda} \ln \lambda}$ ). This is a "quantitative" result because it gives a number, not just a "yes."

3. How They Did It: The "Tuning Fork" Analogy

To get these numbers, the authors used a clever trick involving Ackermann's formula.

Imagine the stubborn low-frequency wiggles are a set of tuning forks that are ringing out of tune.

The Decomposition: They separated the "ringing" (low frequencies) from the "fading" (high frequencies).
The Tuning: They used a mathematical recipe (Ackermann's formula) to calculate exactly how to "tune" the feedback hand so that the low-frequency forks stop ringing immediately.
The Calculation: Most people stop at step 2, saying, "Okay, we tuned it." But these authors went further. They analyzed the math of the tuning recipe to see exactly how the "tuning" gets harder as they try to stop the forks faster. They realized that as they try to silence the forks more quickly, the distance between the "natural" ring and the "forced" silence gets smaller, making the math more complex. They calculated exactly how this complexity scales.

4. The Result: A "Smart Hand" with a Price Tag

The main result (Theorem 1.1) says:

We can build a feedback hand that stops the system at any speed you want.
The Catch: The "price tag" (the maximum force the hand needs to exert) grows exponentially with the speed you want.
The Good News: We now have a precise formula for that price tag. We know exactly how much "muscle" the system needs for a given speed.

5. Why This Matters (According to the Paper)

The authors explain that knowing the "price tag" is crucial for two reasons:

Robustness: If the rope is slightly damaged or there is a small wind blowing on it (a "perturbation"), knowing the cost tells us if our hand is strong enough to handle it. If the cost is too high, even a tiny breeze might knock the system out of control.
Null Controllability (The "Stop on a Dime" Trick): Because they know the exact cost, they can create a "piecewise" strategy. Imagine the hand doesn't just push steadily; it pushes hard, then rests, then pushes harder in a specific pattern. This allows them to bring the rope to a complete standstill (zero energy) in a finite amount of time, not just make it fade away slowly.

Summary

Think of this paper as a mechanic who finally figured out the exact fuel consumption chart for a rocket engine. Before, they knew the rocket could go to the moon. Now, they have a precise chart that says, "To get there in 10 days, you need exactly X gallons of fuel. To get there in 5 days, you need Y gallons."

They took a method that was previously just a "proof of concept" and turned it into a precise engineering blueprint, specifically for one-dimensional systems like the Allen–Cahn equation. This blueprint can now be used to design controllers that are not just stable, but optimally efficient and predictable.

Technical Summary: Quantitative Rapid Boundary Stabilization via Modal Decomposition and its Application to the Allen–Cahn Equation

Problem Statement
This paper addresses the problem of quantitative rapid boundary stabilization for the one-dimensional Allen–Cahn equation, a semilinear parabolic partial differential equation (PDE). The specific model considered is:
$y_t = y_{xx} + y - y^3, \quad (t, x) \in (s, +\infty) \times (0, 1)$
with boundary conditions $y(t, 1) = 0$ and a control input acting on the left boundary $y(t, 0) = a(t)$ . To handle regularity issues, the authors introduce an auxiliary scalar control $u(t)$ governing the evolution of the boundary value $a(t)$ via the ODE $a_t = (1 - \pi^2/4)a + u$ . The state space is defined as $H = L^2(0, 1) \times \mathbb{R}$ .

The primary goal is to design a feedback law $u(t) = K_\lambda(y(t), a(t))$ that ensures the energy of the closed-loop system decays exponentially at an arbitrarily prescribed rate $\lambda > 0$ . Crucially, the paper seeks quantitative estimates for the stabilization cost (the constant $M_\lambda$ in the decay estimate $E(t) \leq M_\lambda e^{-\lambda t} E(0)$ ) and the admissible local radius of initial data, explicitly expressing these dependencies on $\lambda$ .

Methodology
The authors employ a modal decomposition approach combined with the frequency Lyapunov method. The methodology proceeds through the following steps:

State Decomposition: The system is decomposed into low-frequency and high-frequency components. The state is transformed using an auxiliary function $\psi(x) = \cos(\frac{\pi}{2}x)$ to handle the non-homogeneous boundary condition, defining a new variable $w(t) = y(t) - a(t)\psi(x)$ .
Finite-Dimensional Control Design:
- The low-frequency modes (up to index $N$ ) are isolated into a finite-dimensional subsystem.
- The number of modes $N$ is selected based on the desired decay rate $\lambda$ (specifically $N \approx \sqrt{2\lambda}/\pi$ ).
- A linear feedback law is constructed for these low-frequency modes using Ackermann's formula to place the poles of the closed-loop matrix at specific locations $\hat{\lambda}_n = -3\lambda - \lfloor\lambda\rfloor \frac{n}{N}$ . This explicit construction replaces the qualitative existence guarantees of standard pole-placement theorems.
High-Frequency Handling: The high-frequency modes ( $n > N$ ) are stabilized by the intrinsic dissipativity of the heat operator, which dominates the nonlinear terms and the coupling from the low-frequency part.
Quantitative Estimation: The core technical contribution lies in deriving explicit bounds for the feedback gain matrix $K$ $K$ and the Lyapunov matrix $P$ $P$ .
- The authors utilize Ackermann's formula to express $K$ explicitly.
- They estimate the norm of $K$ by analyzing the inverse of a Vandermonde matrix and the product of matrix norms, leading to bounds involving the distances between the original eigenvalues and the assigned pole locations.
- The Lyapunov matrix $P$ is estimated via its integral representation and the spectral properties of the closed-loop matrix $F = A_0 + B_0K$ .
Frequency Lyapunov Function: A composite Lyapunov function $V(t)$ is constructed, balancing the weighted low-frequency energy ( $\gamma Y^\top P Y$ ) and the unweighted high-frequency energy ( $\sum \langle w, \phi_n \rangle^2$ ). The quantitative estimates of $K$ and $P$ allow the authors to prove that $V(t)$ decays exponentially, provided the initial data is sufficiently small.

Key Contributions and Results

Quantitative Rapid Stabilization Theorem (Theorem 1.1):
The paper establishes the existence of a linear feedback law $K_\lambda$ depending only on finite-dimensional modes such that for any $\lambda > 2\pi^2$ , the system admits a unique solution satisfying:
$|u(t)| + \|(y(t), a(t))\|_H \leq e^{D\sqrt{\lambda}\ln \lambda} e^{-\lambda(t-s)} \|(y_0, a_0)\|_H$
for initial data with norm $\|(y_0, a_0)\|_H \leq e^{-D\sqrt{\lambda}\ln \lambda}$ . Here, $D \geq 1$ is a constant. This provides an explicit dependence of the stabilization cost and the admissible initial radius on the decay rate $\lambda$ .
Explicit Feedback Construction:
Unlike traditional modal decomposition which often relies on the qualitative existence of stabilizing gains, this work provides an explicit formula for the feedback gains via Ackermann's formula. This explicitness is essential for deriving the quantitative bounds.
Null Controllability and Finite-Time Stabilization (Corollaries 1.3 & 1.4):
By combining the quantitative rapid stabilization result with piecewise feedback laws and time-iteration arguments, the authors derive:
- Null Controllability: The system can be driven to zero in finite time $T$ with an explicit control cost bound $\|u\|_{L^\infty} \leq e^{Q_p/T^p} \|(y_0, a_0)\|_H$ .
- Finite-Time Stabilization: A periodic piecewise feedback law is constructed that drives the system to zero in time $2T$ for small initial data, ensuring uniform stability.

Significance and Claims
The authors claim this work represents the first quantitative result on rapid stabilization within the modal-decomposition framework.

Novelty 1: The paper overcomes the limitation of standard modal decomposition (which typically yields only qualitative results) by using explicit expressions for feedback gains (Ackermann's formula) and matrix estimation techniques. This allows for the derivation of quantitative low-frequency estimates, specifically reducing the analysis to estimating the distances between pole-placement eigenvalues and the linearized system's eigenvalues.
Novelty 2: The work provides a quantitative realization of the frequency Lyapunov method for boundary rapid stabilization of one-dimensional nonlinear parabolic equations. It extends the qualitative framework established in previous works (e.g., [37]) to a setting where explicit cost estimates are available.
Broader Impact: The authors note that while the results are demonstrated on the Allen–Cahn equation, the quantitative framework is adaptable to a broader class of one-dimensional parabolic models, including those with state delays, stochastic terms, and observer-based controls.

Cost Analysis
The paper highlights that the stabilization cost behaves as $e^{O(\sqrt{\lambda}\ln \lambda)}$ . The authors note that while this is slightly larger than the $e^{O(\sqrt{\lambda})}$ cost found in backstepping approaches (e.g., [5, 35]), the logarithmic factor arises naturally from the estimation of the distances between eigenvalues in the pole-placement procedure. They further observe that for any $p > 1$ , this cost can be bounded by $e^{D\lambda^{\frac{p}{p+1}}}$ , which is sufficient for the time-iteration arguments used to prove null controllability.