Hölder Regularity of Dirichlet Problem For The Complex Monge-Ampère Equation

This paper establishes the global Hölder continuity of solutions to the Dirichlet problem for the complex Monge-Ampère equation on strictly pseudoconvex domains or Hermitian manifolds, assuming the right-hand side belongs to $L^p$ and the boundary data are Hölder continuous.

The Gist

Imagine you are trying to smooth out a crumpled piece of paper (or a bumpy landscape) to make it perfectly flat, but you have two very specific rules to follow:

1. The "Bumpy" Rule: The paper has some random, jagged bumps on it (mathematically, this is the "right-hand side" of the equation). These bumps are messy and not perfectly smooth; they are just "roughly" defined.
2. The "Edge" Rule: The edges of the paper are held down by a frame. The frame itself isn't perfectly smooth either; it has a bit of a rough texture (mathematically, this is the "boundary data").

The Complex Monge-Ampère Equation is the mathematical recipe for finding the smoothest possible shape that fits inside this frame while respecting the bumps underneath.

The Big Question

For a long time, mathematicians knew that if the bumps were very smooth and the frame was very smooth, the resulting shape would be smooth. But what if the bumps are rough (like sandpaper) and the frame is only slightly rough (like a slightly jagged edge)?

Would the final shape be smooth? Or would it be so jagged that it tears? Specifically, the authors wanted to know: Is the final shape "Hölder continuous"?

In plain English, "Hölder continuous" is a fancy way of asking: "If I move a tiny step on the surface, how much does the height change?"

• If the surface is smooth, a tiny step means a tiny height change.
• If the surface is jagged, a tiny step could mean a huge, sudden jump.
• The authors wanted to prove that even with rough inputs, the output doesn't jump wildly; it changes in a controlled, predictable way.

The Old Way vs. The New Way

The Old Approach (The "Heavy Lifting" Method):
Previous mathematicians tried to solve this by building a massive "barrier" (like a giant dam) to hold back the roughness. They calculated the total "mass" or weight of all the bumps inside the domain to estimate the smoothness. It was like trying to smooth a rug by weighing every single fiber. It worked, but the math was heavy, and the resulting smoothness guarantee wasn't the best possible.

The New Approach (Hu and Zhou's "Smart Scaffolding"):
Hu and Zhou, the authors of this paper, came up with a clever, lighter strategy.

1. The "Local Smoothing" Trick: Instead of weighing the whole mountain of bumps, they used a technique called regularization. Imagine taking a blurry camera photo of the surface. If you zoom out a little (blur the image), the tiny jagged spikes disappear, and you see a smooth average. They did this mathematically: they created a "smoothed version" of the solution and compared it to the real solution.
2. The "Edge-First" Strategy: They realized that the most important part of the problem is the edge (the frame). They built a new, smarter "barrier" right at the boundary. Instead of a generic dam, they built a custom-fitted glove that hugs the rough edge perfectly. This allowed them to control the roughness right where it starts.
3. The "Domino Effect": Once they proved the edge was smooth enough, they showed that this smoothness "rippled" inward. Because the edge is controlled, the inside can't get too crazy.

The Result: A Smoother World

The authors proved that even if your "bumps" are rough (mathematically, in an $L^p$ space) and your "frame" is only slightly rough (Hölder continuous), the final shape is guaranteed to be smooth in a specific, measurable way.

They found a new "smoothness score" (the Hölder exponent).

• Previous Score: "It's smooth, but maybe a little bit jagged."
• New Score: "It's significantly smoother than we thought!"

Why Does This Matter?

This isn't just about abstract math. The Complex Monge-Ampère equation is the engine behind Calabi-Yau manifolds, which are the shapes used in String Theory to describe the hidden dimensions of our universe.

• Analogy: If you are trying to build a model of the universe, you need to know if the "fabric" of space is smooth or if it has microscopic tears.
• The Impact: By proving that the solution is smoother than previously thought, Hu and Zhou are telling physicists and mathematicians: "Don't worry, even if the inputs are messy, the geometry of the universe (or the complex shapes we study) remains stable and well-behaved."

Summary

Think of this paper as a master chef who figured out a new way to bake a cake.

• Old Method: You had to use perfect ingredients and a perfect oven to get a smooth cake. If the ingredients were slightly off, the cake might be lumpy.
• New Method: Hu and Zhou discovered a new mixing technique (the barrier and regularization) that guarantees a perfectly smooth cake, even if your ingredients are a little rough and your oven isn't perfect.

They didn't just say "it works"; they gave a precise recipe for how smooth the cake will be, improving on all previous recipes.

Technical Summary

Here is a detailed technical summary of the paper "Hölder Regularity of Dirichlet Problem for the Complex Monge-Ampère Equation" by Yuxuan Hu and Bin Zhou.

1. Problem Statement

The paper addresses the Dirichlet problem for the complex Monge-Ampère (CMA) equation on a strictly pseudo-convex domain $\Omega$ within a complex manifold (either $\mathbb{C}^n$ or a complete Hermitian manifold). The equation is given by:
$$\begin{cases} (dd^c u)^n = f \, d\mu & \text{in } \Omega, \\ u = \phi & \text{on } \partial\Omega, \end{cases}$$
where:

• $u$ is the unknown plurisubharmonic (PSH) function.
• $f \in L^p(\Omega)$ with $p > 1$ is the right-hand side density.
• $\phi \in C^\alpha(\partial\Omega)$ is the boundary data, assumed to be Hölder continuous with exponent $\alpha \in (0,1)$.
• $d\mu$ is the Lebesgue measure (or volume form $\omega^n$ on manifolds).

The Core Question: While the existence of continuous solutions for $f \in L^p$ ($p>1$) was established by Kolodziej, and partial Hölder regularity results existed for smoother boundary data or specific conditions on $f$, the precise global Hölder exponent of the solution $u$ when the boundary data $\phi$ is only $C^\alpha$ (and not necessarily $C^{2\alpha}$ or $C^{1,1}$) remained an open problem with room for improvement. Previous works (e.g., Guedj-Kołodziej-Zeriahi, Charabati) provided exponents that were not optimal, particularly regarding the interplay between the integrability of $f$ and the regularity of $\phi$.

2. Methodology

The authors employ a combination of regularization techniques, barrier constructions, and stability estimates to derive global Hölder continuity. Their approach diverges from previous methods in three key areas:

A. New Barrier Construction for Boundary Regularity

Instead of the traditional decomposition of the solution into a vanishing boundary problem and a homogeneous problem (used in [GKZ, Ch2]), the authors construct a direct barrier function tailored to the boundary geometry.

• They utilize a strictly plurisubharmonic defining function $\rho$ near the boundary (Lemma 3.2).
• By constructing an auxiliary function $h(z) = u(z) + \epsilon + A\rho(z)$ and applying the comparison principle with a solution to a local Dirichlet problem, they derive a sharp boundary Hölder estimate:
  $$|u(x) - u(y)| \leq C \text{dist}(x, y)^\beta \quad \text{for } x \in \Omega, y \in \partial\Omega.$$
• The exponent $\beta$ is optimized to be $\min\{\beta', \frac{\alpha}{2+\alpha}\}$, where $\beta'$ depends on the integrability of $f$. This direct construction allows for a more efficient use of the $L^\infty$-estimate (Theorem 3.1).

B. Elementary $L^1$-Estimate via Regularization

To establish global regularity, the authors analyze the difference between the solution $u$ and its regularization $\hat{u}_\epsilon$ (in $\mathbb{C}^n$) or $\tilde{u}_\epsilon$ (on manifolds).

• Flat Case ($\mathbb{C}^n$): They use standard convolution regularization. A key technical innovation is the estimation of $\|\hat{u}_\epsilon - u\|_{L^1(\Omega_\epsilon)}$. By using Fubini's theorem, they show that the interior contributions cancel out, leaving only the boundary layer contribution. This leads to the estimate:
  $$\|\hat{u}_\epsilon - u\|_{L^1(\Omega_\epsilon)} \leq C \epsilon^{1+\beta}.$$
  This improves upon previous methods that relied on the mass of the Laplacian $\Delta u$.
• Manifold Case: Since standard convolution is not available on manifolds, they use a regularization defined via the exponential map (following [D, BD]). To handle the fact that this regularization is not necessarily PSH, they employ the Kiselman-Legendre transform to construct a PSH regularization $u_{c,\epsilon}$ that satisfies the necessary curvature bounds.

C. Stability Estimates and Iteration

The authors apply the stability estimate for the CMA equation (Theorem 4.1), which bounds the $L^\infty$ difference of two solutions by their $L^r$ difference.

• They combine the boundary estimate (from the barrier) with the $L^1$ estimate of the regularization error.
• Using a generalized modulus of continuity lemma (Lemma 2.1 for flat case, Lemma 2.3 for manifolds), they iterate the estimates to prove that the solution is globally Hölder continuous.

3. Key Contributions and Results

Main Theorem (Theorem 1.1)

The paper proves that the solution $u$ belongs to $C^{\alpha'}(\bar{\Omega})$ with an explicit exponent $\alpha'$:
$$\alpha' = \min\{\beta, (1+\beta)\gamma\}$$
where:

• $\gamma < \gamma_n = \frac{1}{np^* + 1}$ (related to the integrability of $f$).
• $\beta = \max\left\{ \min\left\{\gamma'', \frac{\alpha}{2+\alpha}\right\}, \min\left\{\frac{\alpha}{2}, \gamma'\right\} \right\}$, with $\gamma' < \gamma_n$ and $\gamma'' < \gamma_0 = \frac{1}{p^*+1}$.

Significance of the Exponent:

• The term $\frac{\alpha}{2+\alpha}$ represents a significant improvement over the previous $\frac{\alpha}{2}$ found in works like [Ch2].
• The result holds for complete Hermitian manifolds and extends to complex spaces with isolated singularities (Theorem 4.3).
• The method avoids the requirement that $f$ be bounded near the boundary, a condition previously necessary in some contexts.

Technical Improvements

1. Optimized Boundary Exponent: The new barrier construction yields a boundary Hölder exponent of $\frac{\alpha}{2+\alpha}$ (when $\alpha$ is small), which is strictly better than the $\frac{\alpha}{2}$ obtained by decomposing the problem.
2. Elementary $L^1$ Bound: The proof of the $L^1$ estimate $\|\hat{u}_\epsilon - u\| \leq C\epsilon^{1+\beta}$ is more elementary and direct, avoiding the truncation of the mass of $\Delta u$ used in [BKPZ].
3. Manifold Generalization: The proof successfully adapts to Hermitian manifolds using the Kiselman transform, a non-trivial extension from the Euclidean setting.

4. Significance

This paper resolves a critical gap in the regularity theory of the complex Monge-Ampère equation:

• Optimality: It provides the best known Hölder exponents for solutions with $L^p$ right-hand sides and Hölder continuous boundary data. The improvement from $\alpha/2$ to $\alpha/(2+\alpha)$ is substantial, especially for small $\alpha$.
• Generality: The results apply to a broad class of geometric settings, including non-Kähler Hermitian manifolds and singular complex spaces, making the theory more robust for applications in complex geometry and mathematical physics.
• Methodological Shift: By introducing a direct barrier construction and a refined $L^1$ estimate, the authors provide a new toolkit for analyzing regularity that may be applicable to other fully nonlinear elliptic equations.

In summary, Hu and Zhou establish that the global Hölder regularity of the CMA solution is determined by a delicate balance between the integrability of the source term $f$ and the Hölder continuity of the boundary data $\phi$, achieving a sharper exponent than previously known through innovative barrier and regularization techniques.