Schauder estimates for flat solutions to a class of fully nonlinear elliptic PDEs with Dini continuous data: a geometric tangential approach

This paper establishes local Schauder estimates for flat viscosity solutions to a class of non-convex fully nonlinear elliptic PDEs with Dini continuous data and linear drift terms using geometric tangential techniques, while also deriving an Evans-Krylov type estimate and characterizing the nodal sets of such solutions.

The Gist

Imagine you are trying to predict the weather. You have a massive, complex computer model (the equation) that tries to describe how air, heat, and pressure interact. In the real world, the data you feed into this model—temperature readings, wind speeds, humidity—is never perfect. It's a bit "fuzzy" or "jittery."

For decades, mathematicians have known that if your data is perfectly smooth (like a polished marble), the weather model gives you a perfect, smooth prediction. But what if your data is a bit rough? What if it's "Dini continuous"? That's a fancy math way of saying the data is rough, but not too rough—it's rough in a very specific, controlled way where the "jitter" gets smaller fast enough as you zoom in.

This paper is about proving that even with this slightly rough, "jittery" data, we can still get a very smooth, reliable prediction for a specific type of complex equation.

Here is the breakdown of the paper's story, using some everyday analogies:

1. The Problem: The "Flat" Solution

The authors are looking at a specific class of equations called Fully Nonlinear Elliptic PDEs.

• The Analogy: Imagine a trampoline. If you stand in the middle, it sags. The shape it takes is the "solution."
• The Twist: Usually, we assume the trampoline fabric is uniform and the springs are perfect (convex/concave). But in this paper, the fabric is weird. It might have weird springs that don't follow simple rules (non-convex).
• The "Flat" Condition: The authors focus on solutions that are "flat." Imagine the trampoline is barely sagging at all. It's almost a flat sheet of paper. The authors prove that if the sag is small enough, the weirdness of the springs doesn't matter as much; the shape will still be smooth.

2. The Data: The "Jittery" Input

In math, we usually like our data to be "Hölder continuous" (smooth like silk). But real-world data is often "Dini continuous" (smooth like a slightly worn piece of silk).

• The Analogy: Think of driving a car.
  • Hölder Data: The road is perfectly paved. You can drive at a steady speed.
  • Dini Data: The road has some bumps, but they get smaller and smaller the further you drive. You can still drive smoothly, but you have to be careful.
  • The Old Problem: Previous math theories said, "If the road has any bumps, we can't guarantee a smooth ride."
  • This Paper's Breakthrough: The authors say, "Actually, if the bumps get small fast enough (Dini condition), and the car is driving very slowly (the 'flat' solution), we can guarantee a smooth ride."

3. The Method: The "Geometric Tangential Approach"

This is the cool part. How do they prove it? They use a technique called Geometric Tangential Analysis.

• The Analogy: Imagine you are looking at a giant, bumpy mountain from far away. It looks chaotic. But if you zoom in on a tiny, flat spot on the mountain, it looks like a flat plane.
• The Strategy:
  1. Zoom In: The authors take their complex, weird equation and "zoom in" on a tiny point.
  2. Simplify: As they zoom in, the weird, non-linear parts of the equation start to look like a simple, straight line (a linear equation). It's like how the Earth looks flat when you stand on it, even though it's a sphere.
  3. The "Tangent": They solve the simple, straight-line version of the problem first.
  4. Step Back: They use the solution from the simple version to build a "staircase" back up to the complex version. They prove that because the simple version is smooth, the complex version (which is just a slightly distorted version of the simple one) must also be smooth.

4. The Result: A New Kind of Smoothness

The paper proves that if you have a "flat" solution to this complex equation with "Dini" data, the solution isn't just smooth; it has a specific type of smoothness called $C^{2, \psi}$.

• Translation: This means the "curvature" of your trampoline (the second derivative) is continuous. You can predict exactly how it bends without any sudden jumps.
• Why it matters: This is a huge upgrade. It extends previous work that only worked for "perfectly smooth" (Hölder) data. Now, mathematicians can handle "slightly rough" data and still get precise answers.

5. The "Byproduct": Nodal Sets

The paper also looks at Nodal Sets.

• The Analogy: Imagine a drumhead vibrating. The "nodal set" is the pattern of lines where the drumhead isn't moving at all (the zero lines).
• The Discovery: The authors show that for these flat solutions, these zero lines aren't messy scribbles. They form beautiful, clean geometric shapes (like smooth curves or surfaces). This helps scientists understand the structure of complex physical systems, like how heat spreads or how fluids flow.

Summary

Think of this paper as a new rule for a video game.

• Old Rule: "If the terrain is perfectly smooth, the character can walk smoothly. If the terrain is bumpy, the character stumbles."
• New Rule (This Paper): "If the character is moving very slowly (flat solution) and the bumps in the terrain are getting smaller fast enough (Dini condition), the character can still walk smoothly, even if the terrain isn't perfect."

The authors used a "zoom-in" strategy (Geometric Tangential Approach) to prove this, showing that even in a messy, non-linear world, order and smoothness can still be found if you look at the right scale.

Technical Summary

Here is a detailed technical summary of the paper "Schauder estimates for flat solutions to a class of fully nonlinear elliptic PDEs with Dini continuous data: a geometric tangential approach" by Junior da Silva Bessa, Jo˜ao Vitor da Silva, and Laura Ospina.

1. Problem Statement

The paper addresses the regularity theory for fully nonlinear, uniformly elliptic partial differential equations (PDEs) of the form:
$$F(D^2u, x) + \langle B(x), Du \rangle = f(x) \quad \text{in } B_1 \subset \mathbb{R}^n$$
where:

• $F$ is a fully nonlinear operator that is not necessarily convex or concave (a significant departure from classical Evans-Krylov theory).
• The data ($F$, $B$, and $f$) satisfy Dini continuity conditions rather than the standard Hölder continuity assumptions.
• The authors focus on "flat" solutions, defined as viscosity solutions with a sufficiently small $L^\infty$ norm ($\|u\|_{L^\infty} \leq \delta_0$).

The Core Challenge:
Classical Schauder estimates typically require Hölder continuous coefficients ($C^{0,\alpha}$) to establish $C^{2,\alpha}$ regularity. However, many physical and geometric problems involve data that are Dini continuous but not Hölder continuous (e.g., functions with logarithmic singularities like $r^\alpha / |\log r|^\beta$). The paper seeks to establish local $C^{2, \psi}$ regularity (where $\psi$ is a specific modulus of continuity derived from the Dini condition) for non-convex operators under these weaker continuity assumptions.

2. Methodology: Geometric Tangential Approach

The authors employ a geometric tangential analysis combined with compactness and perturbation arguments. The strategy involves the following steps:

1. Scaling and Normalization:
   The authors define a family of scaled operators $G_\sigma(X) = \frac{1}{\sigma}F(\sigma X)$. As the scaling parameter $\sigma \to 0$, the nonlinear operator $G_\sigma$ converges to a linear, uniformly elliptic operator (the "tangential equation"):
   $$\lim_{\sigma \to 0} G_\sigma(X) = \text{Tr}(A_0 X) = \sum_{i,j} \frac{\partial F}{\partial X_{ij}}(0) X_{ij}$$
   where $A_0$ is the derivative of $F$ at the origin.

2. Approximation by Harmonic Profiles:
   For "flat" solutions (small norm), the solution $u$ to the original nonlinear equation is shown to be arbitrarily close to a solution $U$ of the linear tangential equation $\text{Tr}(A_0 D^2 U) = 0$. Since the linear equation has constant coefficients, its solutions enjoy high regularity ($C^{2,1}_{loc}$).

3. Iterative Geometric Scheme:
   The proof utilizes an induction argument (Lemma 4.3 and Proposition 5.1). At each step $k$, the solution is approximated by a quadratic polynomial $P_k$ on a shrinking ball $B_{\rho^k}$.

   • The error between the solution and the polynomial is controlled by the modulus of continuity $\tau$.
   • The sequence of quadratic polynomials converges to a limiting polynomial $P_\infty$, establishing the pointwise $C^{2,\psi}$ regularity.

4. Handling Non-Convexity:
   Unlike the Evans-Krylov theorem which relies on the convexity of $F$ to obtain $C^{2,\alpha}$ estimates, this approach bypasses the need for convexity by restricting the analysis to flat solutions. The smallness of the solution norm ensures that the nonlinearity behaves sufficiently like its linearization near the origin.

3. Key Assumptions

The results rely on four structural assumptions (A1–A4):

• (A1) Uniform Ellipticity: $F$ satisfies the Pucci extremal operator bounds with constants $0 < \lambda \leq \Lambda$.
• (A2) $C^1$-Regularity of Nonlinearity: $F$ is differentiable with respect to the Hessian variable $X$, with a modulus of continuity $\omega$.
• (A3) Dini Continuity of Data: The source term $f$, the drift $B$, and the oscillation of $F$ ($\theta_F$) are Dini continuous in an $L^{p_0}$ sense ($p_0 > n$). Specifically, their moduli of continuity $\tau$ satisfy $\int_0^1 \frac{\tau(r)}{r} dr < \infty$.
• (A4) Compatibility Conditions: Technical conditions on the modulus of continuity $\tau$ ensuring it decays sufficiently fast (e.g., $\lim_{s\to 0} s/\tau(s) = 0$) and is compatible with the scaling arguments.

4. Main Results

Theorem 1.5 (Local $C^{2, \text{Dini}}$ Regularity)

If $u$ is a viscosity solution to the equation with $\|u\|_{L^\infty(B_1)} \leq \delta_0$ (where $\delta_0$ is a small universal constant), then $u \in C^{2, \psi}_{loc}(B_1)$.
The regularity modulus is given by:
$$\psi(t) = \tau(t) + \int_0^t \frac{\tau(s)}{s} ds$$
The estimate holds:
$$\|u\|_{C^{2, \psi}(B_{1/2})} \leq C_0 \cdot \delta_0$$
This result extends the work of dos Prazeres and Teixeira (who treated Hölder data) to the broader class of Dini continuous data and includes linear drift terms.

Corollary 1.7 (Evans–Krylov Type Estimate)

If a classical solution $u \in C^2(B_1)$ exists, it automatically belongs to $C^{2, \psi}(B_{1/2})$. This provides a regularity upgrade for classical solutions under Dini data.

Corollary 1.8 (Characterization of Nodal Sets)

The authors apply their regularity results to characterize the nodal sets (zero sets) of flat solutions. They show that the sets where the solution and its derivatives vanish up to order $k$ are locally finite unions of $C^{1, \psi}$ manifolds. This generalizes Han's results on nodal sets to non-convex operators with Dini data.

Theorem 1.9 (Convex Case)

Under additional convexity assumptions on $F$ (specifically related to the Alt–Phillips equation context), the authors recover and sharpen previous results by Kovats. Notably, their estimates are independent of the norms of the solution and data, unlike Kovats' framework, and they incorporate drift terms.

5. Significance and Contributions

1. Extension of Schauder Theory: The paper successfully extends Schauder estimates to non-convex fully nonlinear operators under Dini continuity assumptions. This fills a gap where classical Hölder-based theories fail.
2. Inclusion of Drift Terms: The methodology explicitly handles linear drift terms $\langle B(x), Du \rangle$, which were not addressed in previous similar works on flat solutions.
3. Geometric Tangential Framework: The authors demonstrate the robustness of the geometric tangential approach. By linearizing the operator at the origin and using compactness, they avoid the need for global convexity, relying instead on the "flatness" of the solution.
4. Applications to Nodal Sets: The results provide a rigorous foundation for analyzing the structure of singular sets and nodal sets in non-convex PDEs, which is crucial for free boundary problems and geometric analysis.
5. Generality of Moduli: The framework accommodates moduli of continuity that are strictly weaker than Hölder (e.g., logarithmic corrections), offering a more precise description of regularity for rough data.

In summary, this manuscript provides a rigorous, quantitative regularity theory for a broad class of fully nonlinear elliptic PDEs, bridging the gap between classical Hölder theory and the more complex reality of Dini continuous data and non-convex structures.