Abstract:This article presents several numerical techniques for solving Laplace equation. A numerical FORTRAN solver is developed to solve the 2D laplace equation. The numerical approaches implemented in the solver include Jacobi, Gauss-Siedel, Successive Over Relaxation, Alternating Direct Implicit and Multigrid methods. Detailed comparison between different numerical methods is presented and discussed
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: **Compare the efficiency and performance of different numerical methods in solving the two - dimensional Laplace equation**. Specifically, the author developed a FORTRAN solver to solve the two - dimensional Laplace equation, and through implementing several common iterative methods (such as Jacobi, Gauss - Seidel, Successive Over - Relaxation method (SOR), Alternating Direction Implicit method (ADI) and Multigrid method), made a detailed comparison and discussion of these methods.
### Specific problem description
1. **Physical background**
- The Laplace equation is an elliptic partial differential equation, which describes simple physical problems such as incompressible irrotational fluid flow and steady - state heat conduction in solids.
- The mathematical form of the two - dimensional Laplace equation is:
\[
\frac{\partial^2 \psi}{\partial x^2}+\frac{\partial^2 \psi}{\partial y^2} = 0
\]
2. **Selection of numerical methods**
- The paper mainly focuses on the five - point stencil, which is a commonly used second - order accuracy finite - difference method.
- The formula of the five - point stencil is:
\[
\frac{\psi_{i - 1,j}-2\psi_{i,j}+\psi_{i + 1,j}}{\Delta x^2}+\frac{\psi_{i,j - 1}-2\psi_{i,j}+\psi_{i,j + 1}}{\Delta y^2}=0
\]
3. **Comparison of solution methods**
- **Direct methods**: such as Cramer's rule, Gauss elimination method and Thomas algorithm, are simple but require a long computing time.
- **Iterative methods**: such as Jacobi, Gauss - Seidel, SOR, SLOR, ADI and Multigrid methods, are usually more efficient than direct methods.
4. **Research objectives**
- Compare the convergence speed, computing time and stability of different iterative methods.
- Find the optimal relaxation factor \( w \) to accelerate the convergence of SOR, SLOR and ADI methods.
- Verify the correctness and physical rationality of the solver.
### Conclusion
Based on the experimental results, the author draws the following conclusions:
- The Jacobi method is the slowest and requires 1194 iterations to converge.
- The Gauss - Seidel and non - relaxation SOR methods require approximately 624 iterations, which is about half of that of the Jacobi method.
- The SLORB and SLORA methods require 326 iterations respectively, which are significantly faster than the GS and SOR methods.
- Although the ADI method contains two calculations per step, it has the fewest number of iterations.
- The Multigrid method (Multigrid V - cycle) is the most efficient and only needs 6 V - cycles to converge, with a total CPU time of 1.03 milliseconds.
In short, this paper aims to find the best method for solving the two - dimensional Laplace equation by comparing different numerical methods and verify its effectiveness and efficiency.