Global $C^{1,α}$-Regularity for Musielak-Orlicz Equations in Divergence Form

1. Problem Statement

The paper addresses the regularity theory for bounded generalized solutions of nonlinear elliptic partial differential equations (PDEs) in divergence form:
$$-\text{div} A(x, u, Du) = B(x, u, Du)$$
subject to either Dirichlet ($u = \phi$ on $\partial\Omega$) or Neumann ($A \cdot \nu = C$ on $\partial\Omega$) boundary conditions.

The core challenge lies in the non-standard growth of the nonlinearities. Unlike classical equations with polynomial growth (e.g., $p$-Laplacian), the operator $A$ is controlled by a Musielak-Orlicz function $G(x, t)$. This function depends simultaneously on the spatial variable $x$ and the gradient magnitude $t = |\xi|$, encompassing a vast class of models including:

• Variable exponent spaces ($G(x,t) = t^{p(x)}$).
• Standard Orlicz spaces ($G(x,t) = G(t)$).
• Double-phase functionals ($G(x,t) = t^p + a(x)t^q$).
• Multi-phase and more complex heterogeneous growth structures.

The objective is to establish global $C^{1,\alpha}$-regularity (Hölder continuity of the first derivatives) for bounded solutions, extending previous results that were limited to $C^{0,\alpha}$ (Hölder continuity of the solution itself) or restricted to specific sub-cases like variational minimizers.

2. Methodology

The proof strategy combines modern regularity techniques adapted to the generality of the Musielak-Orlicz framework. The methodology proceeds in three main stages:

A. Structural Assumptions and Preliminaries

The author imposes specific structural conditions on the Musielak-Orlicz function $G(x,t)$ and the coefficients:

• Growth Conditions (G0-G3): $G$ satisfies $\Delta_2$-conditions, bounds on the ratio $t g(x,t)/G(x,t)$ (denoted by $g^-$ and $g^+$), and a log-Hölder type continuity condition regarding the spatial dependence of the growth exponent.
• Operator Conditions (A1-A4): The operator $A$ satisfies strict monotonicity, coercivity, and continuity conditions controlled by $G(x,t)$.
• Boundary Conditions: The boundary $\partial\Omega$ is assumed to be of class $C^{1,\beta_3}$ (or satisfying a geometric condition (A) for Dirichlet problems).

B. Higher Integrability (Gehring-Type Lemma)

A crucial step is establishing higher integrability of the gradient $Du$.

• The author proves that if $u$ is a bounded generalized solution, then $|Du|$ belongs to a slightly higher Musielak-Orlicz space $L^{G(x,t)(1+\delta)}$.
• This is achieved via Caccioppoli-type inequalities and the Sobolev-Poincaré inequality adapted to Musielak-Orlicz spaces.
• For the Neumann problem, a novel reflection technique is employed across the boundary (using cubes instead of balls) to extend the interior higher integrability results to the boundary, a gap previously unaddressed in the literature for variable exponent/Orlicz Neumann problems.

C. Comparison with "Frozen" Coefficient Problems

To upgrade from $C^{0,\alpha}$ to $C^{1,\alpha}$, the paper utilizes a comparison argument:

1. Freezing: At a point $x^*$, the coefficients of the operator are "frozen" (treated as constant), creating a simpler auxiliary problem: $-\text{div} A(x^*, u(x^*), Dv) = 0$.
2. Regularity of the Auxiliary Problem: The solution $v$ to this frozen problem is known to have higher regularity (specifically $C^{1,\lambda}$) due to the uniform ellipticity of the frozen operator.
3. Estimation: The difference between the original solution $u$ and the frozen solution $v$ is estimated. Using the continuity of $A$ with respect to $x$ and $u$, and the higher integrability of $Du$, the author derives Campanato-type estimates for the oscillation of $Du$.
4. Conclusion: These estimates imply that $Du$ is Hölder continuous.

3. Key Contributions

1. Global $C^{1,\alpha}$ Regularity in a Unified Framework: The paper provides the first proof of global $C^{1,\alpha}$ regularity for bounded generalized solutions of elliptic equations with general Musielak-Orlicz growth, covering Dirichlet and Neumann problems simultaneously.
2. Extension Beyond Variational Minimizers: Unlike previous works (e.g., Hästö and Ok) that focused on minimizers of energy functionals, this work addresses generalized solutions to PDEs that may not arise from a variational principle. This requires stronger structural assumptions on the nonlinearity but significantly broadens the applicability of the theory.
3. Neumann Boundary Regularity: The paper establishes higher integrability and $C^{1,\alpha}$ regularity for Neumann problems with non-standard growth, a result not previously available in the literature for such general settings.
4. Refined Structural Conditions: The author introduces specific conditions (G2 and G3) regarding the interplay between the spatial variation of the growth function and the boundary behavior, ensuring the regularity holds even when the growth is highly non-uniform.

4. Main Results

The paper establishes three main theorems under the assumption that the solution $u$ is bounded ($u \in L^\infty(\Omega)$):

• Theorem 1.4 (Interior Regularity): If $u$ is a bounded generalized solution of the equation in the interior of $\Omega$, then $u \in C^{1,\alpha}_{\text{loc}}(\Omega)$. The Hölder exponent $\alpha$ depends on the dimension, growth bounds ($g^-, g^+$), and structural constants.
• Theorem 1.5 (Dirichlet Boundary Regularity): If $\partial\Omega$ is $C^{1,\beta_3}$ and $u$ satisfies Dirichlet boundary conditions with $C^{1,\beta_3}$ data, then $u \in C^{1,\alpha}(\overline{\Omega})$.
• Theorem 1.6 (Neumann Boundary Regularity): If $\partial\Omega$ is $C^{1,\beta_3}$ and $u$ satisfies Neumann boundary conditions with appropriate data regularity, then $u \in C^{1,\alpha}(\overline{\Omega})$.

In all cases, the Hölder exponent $\alpha$ and the norm $\|u\|_{C^{1,\alpha}}$ depend explicitly on the structural parameters ($n, g^-, g^+, \lambda_0, \Lambda_0, F, M$, etc.) but are independent of the specific solution $u$ (provided it is bounded).

5. Significance

• Unification: The results unify and generalize numerous existing theories. They recover known results for $p(x)$-Laplacians, double-phase problems, and standard Orlicz equations as special cases, providing a single analytical framework for all.
• Theoretical Advancement: By moving from $C^{0,\alpha}$ to $C^{1,\alpha}$ in the Musielak-Orlicz setting, the paper bridges a critical gap in the regularity theory of non-standard growth PDEs. This is essential for applications where the smoothness of the gradient is required (e.g., in fluid dynamics, elasticity, and image processing).
• Robustness: The methodology demonstrates that despite the complexity of spatially dependent, non-uniform growth, the fundamental mechanisms of elliptic regularity (higher integrability and comparison principles) remain robust, provided the growth function satisfies specific continuity and structural constraints.

In summary, this paper represents a significant step forward in the analysis of nonlinear elliptic equations with non-standard growth, offering a comprehensive regularity theory that applies to a wide spectrum of physical and mathematical models involving heterogeneous media.