Abstract
The Swift–Hohenberg equation is a widely studied fourth-order model, originally proposed to describe hydrodynamic fluctuations. It admits an energy-dissipation law and, under suitable assumptions, bounded solutions. Many structure-preserving numerical schemes have been proposed to retain such properties; however, existing approaches are often fully implicit and therefore computationally expensive. We introduce a simple design principle for constructing dissipation-preserving finite difference schemes and apply it to the Swift–Hohenberg equation in three spatial dimensions. Our analysis relies on discrete inequalities for the underlying energy, assuming a Lipschitz continuous gradient and either convexity or $\mu$-strong convexity of the relevant terms. The resulting method is linearly implicit, yet it preserves the original energy-dissipation law, guarantees unique solvability, ensures boundedness of numerical solutions, and admits an a priori error estimate, provided that the time step is sufficiently small. To the best of our knowledge, this is the first linearly implicit finite difference scheme for the Swift–Hohenberg equation for which all of these properties are established.
Introduction
The Swift–Hohenberg (SH) equation is a cornerstone of pattern formation physics. Originally derived to model hydrodynamic fluctuations, its applications have expanded into fields as diverse as optical fibers and neuroscience. Mathematically, the SH equation is a fourth-order partial differential equation (PDE) characterized by a crucial physical property: an energy-dissipation law. This law ensures that the system’s total energy non-increases over time, leading to the formation of stable, complex patterns.
To simulate these systems accurately, researchers utilize "structure-preserving" numerical schemes. However, a significant trade-off has historically existed between computational efficiency and mathematical reliability. Most schemes that guarantee stability are fully implicit, requiring the solution of complex, nonlinear systems at every time step. While faster "linearly implicit" schemes exist, they often lack the theoretical guarantees necessary to ensure that solutions remain bounded or strictly follow the energy-dissipation law. This paper introduces a new splitting principle that bridges this gap.
Methodology: Smooth-Convex-Concave Splitting
The authors propose a design principle termed the Smooth-Convex-Concave Splitting Scheme. This approach decomposes the energy functional of the SH equation into three distinct components:
- Smooth: Terms characterized by a Lipschitz continuous gradient.
- Convex: Terms that exhibit a "curving upward" property.
- Concave: Terms that exhibit a "curving downward" property.
By treating the convex components implicitly (incorporating information from the next time step) and the smooth and concave components explicitly (using current data), the researchers transform the difficult nonlinear problem into a linear system. This allows for significantly faster computation while maintaining the structural integrity of the underlying physics.
The analysis is conducted within a finite difference framework in three spatial dimensions. To ensure the scheme's robustness, the authors leverage tools from convex analysis, specifically $\mu$-strong convexity, to guarantee that a unique solution exists at every discrete time step.
Key Theoretical Insights
The Power of Strategic Splitting
The research demonstrates that it is not necessary to make an entire equation implicit to maintain stability. By strategically splitting the right-hand side of the PDE, the scheme preserves the original energy-dissipation law while requiring only the solution of linear equations. This significantly reduces the computational overhead per time step.
A Posteriori Boundedness
A major contribution of this work is the proof of boundedness. Using a modified functional and mathematical induction, the authors prove that if the initial state of the system is bounded, the numerical solution will remain bounded for all subsequent time steps, provided the time step ($\Delta t$) is chosen appropriately. This prevents the numerical "blow-up" common in less rigorous schemes.
Mesh-Independent Stability
Crucially, the restriction on the time step ($\Delta t$) depends on the physical constants of the equation (such as $\eta$) but remains independent of the spatial mesh size ($\Delta x$). This allows researchers to refine the spatial grid for higher resolution without being forced to use an infinitesimally small time step, a common limitation in many numerical methods for fourth-order PDEs.
Results and Convergence
The paper provides a rigorous mathematical foundation for the proposed scheme, establishing several critical properties:
- Unique Solvability: The scheme is guaranteed to produce exactly one solution at every time step.
- Energy Dissipation: The discrete energy is proven to be non-increasing, mirroring the physical reality of the SH equation.
- Convergence Rate: The authors derive an a priori error estimate showing the method is first-order accurate in time and second-order accurate in space, converging at a rate of $O(\Delta t + (\Delta x)^2 + (\Delta y)^2 + (\Delta z)^2)$.
- Stability: The numerical solution is guaranteed not to diverge to infinity.
Conclusion and Future Directions
This work represents the first linearly implicit finite difference scheme for the Swift–Hohenberg equation to simultaneously satisfy unique solvability, energy dissipation, boundedness, and convergence. While the current scheme is first-order accurate in time, future research may extend this framework to second-order accuracy (e.g., via Crank-Nicolson methods) while maintaining these safety guarantees. Additionally, adapting these proofs for more complex boundary conditions beyond Neumann and periodic boundaries remains an open area for exploration.
Summary updated