🔢 mathematics

Maximal regularity for evolution equations with critical singular perturbations

This paper establishes maximal $L^p$ -regularity for evolution equations perturbed by time-dependent unbounded operators satisfying critical $L^q$ -integrability conditions, extending previous results to the critical endpoint and developing a weighted theory for mixed-scale perturbations relevant to stochastic PDEs.

Original authors: Esmée Theewis, Mark Veraar

Published 2026-02-27

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Esmée Theewis, Mark Veraar

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 bake a perfect cake (solving a complex mathematical equation) in a very specific, high-tech kitchen. The recipe you have is for a standard cake, and you know exactly how to bake it perfectly every time. In math terms, this "perfect cake" is called Maximal Regularity. It means you have a reliable method to predict exactly how the ingredients (the solution) will behave over time, ensuring the cake doesn't collapse or turn into a mess.

However, in the real world (and in advanced physics), things aren't always perfect. Sometimes, you have to add a "special ingredient" to the batter. This ingredient is singular (it's very potent and unstable) and time-dependent (it changes strength as you mix).

This paper is about figuring out: "If I add this wild, unstable ingredient to my perfect recipe, will the cake still turn out okay?"

Here is the breakdown of their discovery using everyday analogies:

1. The "Perfect Recipe" (The Operator A)

Think of the main part of your equation, the operator A, as the standard oven and mixing process. Mathematicians have known for a long time that if you just use the oven and mix the flour and eggs (the standard parts of the equation), you get a perfect result. The cake rises exactly as predicted.

2. The "Wild Ingredient" (The Perturbation B)

Now, imagine you want to add a new spice, B.

The Problem: This spice is "singular." It's like a spice that is infinitely strong for a split second and then disappears, or a spice that is so potent it could burn the cake if you aren't careful.
The Critical Condition: The authors look at a specific "critical" case. Imagine the spice is added at a rate that is just barely safe. It's on the razor's edge. If you add a tiny bit more, the cake burns. If you add a tiny bit less, the flavor is lost. This is the "Critical Singularity."

3. The First Discovery: The "Edge-of-Chaos" Rule

The first major result of the paper is like discovering a new rule for cooking:

"Even if this wild spice is added at the absolute maximum limit of safety (the critical endpoint), as long as you measure it correctly over time, the cake will still bake perfectly."

Previously, mathematicians thought that if the spice was this strong, the cake would inevitably fail. The authors proved that with the right mathematical "technique" (using specific weights and measuring the spice's intensity in a specific way), the cake holds together. They generalized a previous rule that only worked for "safer" spices to now include these "dangerous" ones.

4. The Second Discovery: The "Mixed-Texture" Cake

The second part of the paper is even more complex. It's motivated by Stochastic Partial Differential Equations (SPDEs).

The Analogy: Imagine you are baking a cake, but the kitchen is shaking (random noise, like a storm outside). To predict the cake's behavior in a storm, you need to look at a "Skeleton Equation." This is like a simplified blueprint of the cake that ignores the shaking but keeps the core structure.
The Challenge: In this blueprint, the "wild spice" (B) isn't just added to the main batter; it's added to a different layer of the cake (a different mathematical space). It's like adding the spice to the frosting and the sponge at the same time, but the frosting and sponge react differently.
The Solution: The authors created a new way to measure the cake's quality. Instead of just checking if the cake is "perfectly smooth" (standard regularity), they developed a "Mixed-Scale" check.
- They check the cake's structure in two ways simultaneously:
  1. Is the main sponge solid? (Standard check).
  2. Is the frosting layer holding up, even though it's being shaken? (The new, specialized check).
- They proved that even with this double-layered, chaotic spice, the cake remains stable.

5. Why Does This Matter? (The "Why Should I Care?")

You might ask, "Who cares about math cakes?"

This research is crucial for Large Deviation Theory.

Real World Analogy: Imagine you are a risk manager for a bank. You want to know the odds of a "black swan" event—a catastrophic crash that happens once in a million years.
To calculate these tiny, dangerous odds, you have to look at the "Skeleton" of the financial system (the equation mentioned above).
The "wild spice" in the equation represents the extreme, rare fluctuations in the market.
The Impact: Before this paper, mathematicians could only calculate these risks in very simple, safe scenarios. This paper gives them the tools to calculate risks in complex, real-world scenarios where the market is volatile, the data is messy, and the "spice" is dangerously strong. It allows scientists to predict rare disasters in fields like finance, climate science, and fluid dynamics with much higher confidence.

Summary

The Problem: Can we solve complex equations when we add a "dangerous," unstable ingredient that is on the very edge of being too strong?
The Answer: Yes!
The Method: They developed new mathematical "safety nets" (weighted spaces and mixed-scale checks) that prove the system remains stable even at the breaking point.
The Result: We can now model and predict rare, extreme events in the real world (like financial crashes or extreme weather) much more accurately than before.

1. Problem Statement and Motivation

The paper addresses the maximal $L^p$ -regularity of non-autonomous linear evolution equations subject to critical, time-dependent singular perturbations. The primary equation of interest is:
$u'(t) + A(t)u(t) + B(t)u(t) = f(t), \quad u(0) = u_0$
where:

$A(t)$ is a time-dependent operator family on a Banach space $X_0$ with domain $X_1$ .
$B(t)$ is a lower-order perturbation that is singular in time (unbounded in norm) and satisfies a critical integrability condition.
The perturbation is "critical" because the space-time Sobolev indices of the leading term $Au$ and the perturbation $Bu$ coincide. This places $B$ on the borderline of standard perturbation theories, making it non-trivial to prove that the perturbed operator $A+B$ retains maximal regularity.

Motivation:
The work is driven by Large Deviations Theory (LDT) for Stochastic Partial Differential Equations (SPDEs). In the weak convergence approach to LDT, one analyzes "skeleton equations" where a multiplicative noise term $B(t)u(t)dW(t)$ is replaced by a deterministic control term $B(t)\phi(t)dt$ .

In the $L^2$ setting, previous work (e.g., [28]) handled cases where $A$ was coercive.
This paper aims to extend these results to $L^p$ settings with $p \in (2, \infty)$ and to handle non-coercive operators $A$ , which are common in general SPDE applications.
The specific challenge is that the control function $\phi$ typically belongs to $L^2$ , making the perturbation $B(t)\phi(t)$ belong to a critical $L^r$ space where $1/r = 1/p + 1/2$ .

2. Methodology

The authors develop a new perturbation theory that avoids standard methods like the "method of continuity" (which fails for singular perturbations) and instead employs a gluing procedure (iterative construction on small time intervals) combined with weighted maximal regularity.

Key Technical Tools:

Power-Weighted Spaces: The authors utilize spaces $L^p(0, T, w_\kappa; X)$ with weights $w_\kappa(t) = t^\kappa$ . This allows for rough initial data $u_0 \in X_{1-1/p, p}$ and handles the singularity at $t=0$ .
Interpolation Theory: Extensive use of real and complex interpolation spaces (e.g., $X_{1-1/q, 1}$ , $X_{1-1/p, p}$ ) and trace theorems to characterize the regularity of solutions and the domain of the perturbation.
Gluing/Iterative Strategy:
- Instead of solving the global problem directly, the authors construct local solutions on small time intervals $[0, \tau]$ .
- They choose $\tau$ such that the $L^q$ -norm of the perturbation bound $b(t)$ is sufficiently small to ensure the perturbation map is a contraction.
- Solutions are then "glued" together iteratively to cover the interval $[0, T]$ .
- Crucially, they prove that the constants in the estimates depend only on the $L^q$ -norm of the perturbation bound, not on the specific shape of the function, even though the number of iterations depends on this norm.
Mixed-Scale Regularity: For the skeleton equation application, the solution space is not a standard maximal regularity space but a sum space:
$\mathcal{MR}_+ = \left( L^p \cap W^{1,p} \right) + \left( L^r \cap W^{1,r} \right)$
This accommodates the fact that the inhomogeneity $Bu$ may not lie in $L^p(X_0)$ but in a lower regularity space $L^r(X_\gamma)$ .

3. Key Contributions and Results

The paper establishes two main theorems corresponding to two different regimes of perturbation.

Result 1: General Critical Perturbations (Theorem 2.5 & 2.6)

Setting: $A$ has maximal $L^p$ -regularity. $B(t)$ maps $X_{1-1/q, 1} \to X_0$ with $\|B(t)\| \leq b(t)$ where $b \in L^q(0, T)$ .
Condition: $q \in (p, \infty)$ (Theorem 2.5) or the limiting case $q=p$ (Theorem 2.6).
Contribution: This generalizes the result of Prüss and Schnaubelt (2001) by addressing the critical endpoint case ( $b \in L^q$ rather than $L^{q+\epsilon}$ ).
Outcome: The perturbed operator $A+B$ retains maximal $L^p$ -regularity. The solution $u$ belongs to the standard maximal regularity space $L^p(0, T; X_1) \cap W^{1,p}(0, T; X_0)$ .

Result 2: Mixed-Scale Perturbations for Skeleton Equations (Theorem 3.11 & 4.3)

Setting: Designed specifically for the skeleton equation $u' + Au + Bu = f + g$.
- $B(t): X_1 \to X_\gamma$ (where $\gamma \in (0, 1)$ ).
- $B(t)$ satisfies a critical integrability condition related to $L^2$ controls (e.g., $b \in L^2$ ).
- The inhomogeneity $f$ is in a sum space $L^p(X_0) + L^r(X_\gamma)$ .
Main Theorem (3.11): Proves well-posedness in a mixed-scale maximal regularity space $\mathcal{MR}_+$ $M R_{+}$ .
- The solution $u$ satisfies $u \in L^p(0, T; X_1) \cap C([0, T]; X_{1-1/p, p})$ .
- The theory handles non-coercive operators $A$ and time-dependent coefficients.
Limiting Case $r=1$ (Theorem 4.3): Addresses the case $p=2$ (where $r=1$ ), which corresponds to the $L^2$ -setting with $L^1$ perturbations. This recovers and extends previous variational results to a non-variational setting using a transference method involving an auxiliary operator $A_0$ generating an analytic semigroup.

4. Significance and Impact

Resolution of Critical Endpoint Problems: The paper successfully handles the "borderline" integrability conditions ( $L^q$ with $q$ exactly matching the critical exponent) which were previously out of reach for standard perturbation techniques.
Advancement in Large Deviations Theory: By establishing $L^p$ -maximal regularity for the linearized skeleton equations with $p > 2$ , the authors provide the necessary analytical foundation to extend Large Deviation Principles to a broader class of SPDEs, particularly those where the leading operator is not coercive or where $L^2$ -based variational methods are insufficient.
Robustness to Singularities: The developed framework allows for time-dependent operators $A(t)$ that are unbounded in time and perturbations $B(t)$ that are singular, provided they satisfy specific integrability constraints.
Generalization of Existing Theory: The results generalize the work of Prüss, Schnaubelt, and the authors' own previous work, unifying the treatment of autonomous and non-autonomous cases, and covering both coercive and non-coercive scenarios.

Summary of Technical Table (Table 1)

The paper categorizes results based on the perturbation $B$ and the resulting solution space:

Standard Perturbation: $B \in L^q(X_{1-1/q} \to X_0)$ with $q > p$ . Solution is in standard $MR_p$ .
Critical Perturbation: $B \in L^p(X_{1-1/p} \to X_0)$ . Solution is in standard $MR_p$ .
Mixed-Scale (Skeleton): $B \in L^2(X_1 \to X_{1/2})$ with $f \in L^p + L^r$ . Solution is in the sum space $\mathcal{MR}_+$ , embedding into $L^p(X_1) \cap C(X_{1-1/p})$ .
Limiting Case ( $r=1$ ): $B \in L^{p'}(X_1 \to X_{1-1/p})$ . Solution in $L^p(X_1) \cap C(X_{1-1/p})$ .

In conclusion, this paper provides a rigorous and flexible analytical framework for evolution equations with critical singular perturbations, directly enabling new applications in the stochastic analysis of SPDEs via Large Deviations Theory.