Abstract:The Neumann--Neumann method is a commonly employed domain decomposition method for linear elliptic equations. However, the method exhibits slow convergence when applied to semilinear equations and does not seem to converge at all for certain quasilinear equations. We therefore propose two modified Neumann--Neumann methods that have better convergence properties and require less computations. We provide numerical results that show the advantages of these methods when applied to both semilinear and quasilinear equations. We also prove linear convergence with mesh-independent error reduction under certain assumptions on the equation. The analysis is carried out on general Lipschitz domains and relies on the theory of nonlinear Steklov--Poincaré operators.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: The existing Neumann - Neumann methods show problems of slow convergence or no convergence at all when applied to semilinear and quasilinear elliptic equations. Specifically, for some quasilinear equations, the standard Neumann - Neumann method cannot provide an effective numerical solution.
To solve these problems, the author proposes two improved Neumann - Neumann methods (MNN1 and MNN2), which have better convergence performance and require less computation. The main objectives of the paper include:
1. **Provide numerical evidence**: Demonstrate the advantages of the improved Neumann - Neumann methods in dealing with semilinear and degenerate quasilinear elliptic equations.
2. **Theoretical proof**: Rigorously prove the linear convergence of the improved methods on linear and quasilinear elliptic equations, especially in the case where only the domain and sub - domains are assumed to be Lipschitz conditions.
Through these improvements, the author hopes to significantly improve the convergence speed and stability of solving semilinear and quasilinear elliptic equations while maintaining efficient parallel computing.
### Summary of Mathematical Formulas
The mathematical formulas involved in the paper mainly include the following types:
1. **Quasilinear elliptic equation**:
\[
\begin{cases}
-\nabla \cdot \alpha(x, u, \nabla u) + \beta(x, u, \nabla u) = f & \text{in } \Omega, \\
u = 0 & \text{on } \partial \Omega.
\end{cases}
\]
2. **Weak form**:
\[
\langle Au, v \rangle = \int_{\Omega} \alpha(x, u, \nabla u) \cdot \nabla v + \beta(x, u, \nabla u) v \, dx,
\]
where \( A: V \to V^* \) is a nonlinear operator.
3. **Steklov - Poincaré operator**:
\[
\langle S_i \eta, \mu \rangle_{\Lambda^* \times \Lambda} = \langle A_i F_i \eta - f_i, R_i \mu \rangle_{V_i^* \times V_i},
\]
where \( F_i: \Lambda \to V_i \) is a solution operator.
4. **Improved iterative format**:
- MNN1:
\[
\eta^{n + 1} = \eta^n - s P^{-1} S \eta^n,
\]
where \( P^{-1} = s_1 P_1^{-1} + s_2 P_2^{-1} \).
- MNN2:
\[
\eta^{n + 1} = \eta^n - s P(\eta^n)^{-1} S \eta^n,
\]
where \( P(\nu)^{-1} = s_1 P_1(\nu)^{-1} + s_2 P_2(\nu)^{-1} \).
Through these improvements, the paper shows the effectiveness and superiority of the improved Neumann - Neumann methods in dealing with complex elliptic equations.