Analysis of the Vanishing Moment Method and its Finite Element Approximations for Second-order Linear Elliptic PDEs in Non-divergence Form

Xiaobing Feng, Thomas Lewis, Stefan Schnake

Abstract

This paper investigates the continuous and discrete approximations of $W^{2,p}$ strong solutions for second-order linear elliptic partial differential equations (PDEs) in non-divergence form. By employing the Vanishing Moment Method (VMM), which introduces a small biharmonic regularization term, the authors transform the problem into a fourth-order PDE suitable for Galerkin-type methods. The study establishes the well-posedness of the weak form, provides error estimates for the strong solution approximation, and proposes a $C^1$ finite element method. Numerical results validate the theoretical findings, demonstrating optimal convergence rates and stability even in degenerate cases.

Introduction

Second-order linear elliptic PDEs in non-divergence form, typically expressed as $-A : D^2u + \mathbf{b} \cdot \nabla u + cu = f$, are prevalent in fields such as stochastic control, mathematical finance, and game theory. Unlike equations in divergence form (like the Poisson equation), these PDEs lack a natural weak structure because the coefficient matrix $A$ cannot be easily moved outside the second derivative via integration by parts. Consequently, developing stable and convergent numerical schemes for these "strong solutions" has historically been a significant challenge.

This research addresses this gap by utilizing the Vanishing Moment Method (VMM). By perturbing the original equation with a higher-order term, the authors create a framework that allows for standard Galerkin-based numerical analysis while maintaining a rigorous connection to the original second-order problem.

The Vanishing Moment Method (VMM)

The core idea of the VMM is to regularize the second-order PDE by adding a small biharmonic term: $$\epsilon \Delta^2 u^\epsilon - A : D^2 u^\epsilon = f$$ where $\epsilon > 0$ is a small parameter. This transformation serves several critical purposes:

  1. Divergence Form Transformation: The biharmonic term $\Delta^2 u$ is in divergence form, which facilitates the use of integration by parts. This converts a problem that is traditionally difficult to discretize into one that fits perfectly into the Galerkin framework.
  2. Regularization and Regularity: The higher-order term acts as a bridge, smoothing the problem. The authors show that even if the original solution $u$ is only in $H^2$, the perturbed solution $u^\epsilon$ can exhibit higher regularity (up to $H^4$), depending on the boundary's smoothness.
  3. Boundary Conditions: The authors adopt "simply supported" boundary conditions ($u=0$ and $\Delta u=0$), which naturally absorb into the weak formulation and allow for high-order regularity estimates.

Finite Element Discretization

To solve the resulting fourth-order PDE numerically, the approximation space must be $C^1$ continuous across element boundaries. The authors propose a $C^1$ conforming finite element method using:

  • Argyris Elements: Utilizing 5th-degree polynomials to ensure high-order smoothness.
  • Cubic Hermite Elements: Providing a balance between computational efficiency and regularity.

Through the use of "freezing coefficient" techniques and non-standard duality arguments, the paper proves that this $C^1$ FEM scheme is well-posed and achieves optimal error estimates in the $H^2$ norm.

Key Insights and Theoretical Results

The analysis yields several critical insights regarding the convergence of $u^\epsilon$ to the true solution $u$ as $\epsilon \to 0$:

  • Convergence Rates: The authors proved that the error between the VMM solution and the true solution is bounded by $\sqrt{\epsilon}$ in the $L^2$ and $H^1$ norms.
  • Avoiding the Babuška Paradox: By driving the regularization parameter $\epsilon$ to zero in tandem with the mesh size $h$, the VMM naturally avoids the convergence failures typically seen when approximating curved boundaries for biharmonic equations (the Babuška paradox).
  • Stability: The method remains stable even when the solution has discontinuous second or third-order derivatives, making it a robust choice for non-smooth problems.

Numerical Validation

The method was tested using manufactured solutions in 2D to verify convergence rates across several scenarios:

  • Square Domains: Tests with solutions in $H^3$ and $H^2$ (where derivatives are discontinuous along a line) confirmed the method's stability.
  • Circular Domains: Used to test the method's resilience against boundary approximation issues, confirming the theoretical mitigation of the Babuška paradox.
  • Degenerate Cases: A "hard" problem involving a degenerate elliptic matrix where the solution is only a "viscosity solution" showed that the VMM still converges, hinting at its utility for fully nonlinear equations like Monge-Ampère.

In practice, numerical experiments showed linear convergence ($O(\epsilon)$) in the $L^2$ norm, which is superior to the theoretically proven $\sqrt{\epsilon}$ bound, suggesting that the current theoretical estimates may be conservative.

Conclusion

The Vanishing Moment Method provides a powerful bridge between the theory of strong solutions and practical numerical implementation for non-divergence form PDEs. By transforming these problems into fourth-order equations, the authors have enabled the use of standard $C^1$ finite element tools with proven stability and optimal convergence. This work serves as a crucial starting point for applying VMM to more complex, fully nonlinear second-order PDEs.

Summary updated

Abstract

This paper is concerned with continuous and discrete approximations of $W^{2,p}$ strong solutions of second-order linear elliptic partial differential equations (PDEs) in non-divergence form. The continuous approximation of these equations is achieved through the Vanishing Moment Method (VMM) which adds a small biharmonic term to the PDE. The structure of the new fourth-order PDE is a natural fit for Galerkin-type methods unlike the original second order equation since the highest order term is in divergence form. The well-posedness of the weak form of the perturbed fourth order equation is shown as well as error estimates for approximating the strong solution of the original second-order PDE. A $C^1$ finite element method is then proposed for the fourth order equation, and its existence and uniqueness of solutions as well as optimal error estimates in the $H^2$ norm are shown. Lastly, numerical tests are given to show the validity of the method.
BibTeX
@inproceedings{feng2019analysis,
  author = {Xiaobing Feng and Thomas Lewis and Stefan Schnake},
  title = {{Analysis of the Vanishing Moment Method and its Finite Element Approximations for Second-order Linear Elliptic PDEs in Non-divergence Form}},
  booktitle = {Methods},
  year = {2019},
  doi = {10.4310/maa.2019.v26.n2.a5},
}
Figure