Abstract:This work delves into the exponential time differencing (ETD) schemes for the matrix-valued Allen-Cahn equation. In fact, the maximum bound principle (MBP) for the first- and second-order ETD schemes is presented in a prior publication [SIAM Review, 63(2), 2021], assuming a symmetric initial matrix field. Noteworthy is our novel contribution, demonstrating that the first- and second-order ETD schemes for the matrix-valued Allen-Cahn equation -- both being linear schemes -- unconditionally preserve the MBP, even in instances of nonsymmetric initial conditions. Additionally, we prove that these two ETD schemes preserve the energy dissipation law unconditionally for the matrix-valued Allen-Cahn equation. Some numerical examples are presented to verify our theoretical results and to simulate the evolution of corresponding matrix fields.
What problem does this paper attempt to address?
### What problems does this paper attempt to solve?
This paper mainly explores the exponential time differencing (ETD) method for the matrix - valued Allen - Cahn equation and solves the following key problems:
1. **Preservation of the Maximum Bound Principle (MBP)**:
- For the first - order and second - order ETD schemes, the author proves that these linear schemes can unconditionally preserve the maximum bound principle even when the initial conditions are asymmetric. This is different from previous research, where previous analyses usually required the initial matrix field to be symmetric.
2. **Preservation of the Energy Dissipation Law**:
- The author proves for the first time that the first - order and second - order ETD schemes can unconditionally satisfy the energy dissipation law of the matrix - valued Allen - Cahn equation. This means that the numerical method can ensure that the total energy of the system decreases as time evolves, thus ensuring that the numerical solution is consistent with the evolution behavior of the actual physical system.
3. **Convergence analysis**:
- The paper also provides error estimates for these ETD methods, with coefficients independent of the parameter ε. This is very important for evaluating the accuracy and stability of the numerical method.
### Formula summary
- **Energy function**:
\[
E(U)=\int_{\Omega}\left(\frac{\varepsilon^{2}}{2}\|\nabla U\|_{F}^{2}+\langle F(U), I\rangle_{F}\right)dx
\]
where,
\[
F(U)=\frac{1}{4}(U^{T}U - I)^{2}, \quad f(U)=U - UU^{T}U
\]
- **Maximum bound principle**:
\[
\max_{x\in\Omega}\|U(t, x)\|_{F}\leq\sqrt{m}, \quad t\in(0, T]
\]
- **Energy dissipation law**:
\[
E(U^{n + 1})\leq E(U^{n}), \quad \forall n\geq0
\]
### Research background and significance
The matrix - valued Allen - Cahn equation is widely used in image processing, geometric processing, computer graphics and other fields, and has important applications especially in the study of polycrystalline materials. By improving the existing ETD method, this paper makes the numerical simulation more accurate and efficient, and at the same time relaxes the restrictions on the initial conditions, making it more suitable for general situations. This provides a solid foundation for subsequent theoretical research and practical applications.