Abstract:Finite Difference (FD) schemes are widely used in science and engineering for approximating solutions of partial differential equations (PDEs). Error analysis of FD schemes relies on estimating the truncation error at each time step. This approach usually leads to a global error whose order is of the same order of the truncation error. For classical FD schemes the global error is indeed of the same order as the truncation error. A particular class of FD schemes is the Block Finite Difference (BFD) schemes, in which the grid is divided into blocks. The structure of such schemes is similar to the structure of the Discontinuous Galerkin (DG) method, and allows inhabitation of the truncation errors. Recently, much effort was devoted to design BFD schemes whose global error converges faster than the truncation error. In this paper, we elaborate the approach presented in <a class="link-https" data-arxiv-id="1711.07926" href="https://arxiv.org/abs/1711.07926">arXiv:1711.07926</a> for the heat equation with periodic boundary conditions. We generalize this methodology to design BFD schemes for the heat equation with Dirichlet or Neumann boundary conditions, whose global error converges faster than the truncation error. Such schemes are henceforth called Error Inhibiting Schemes. We provide an explicit error analysis, including proofs of stability and convergence of the proposed schemes. We illustrate our approach using several numerical examples, which demonstrate the efficiency of our method in comparison to standard FD schemes.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to design a new Finite Difference (FD) scheme in order to increase the global error convergence rate of the numerical solution of the heat conduction equation in Initial Boundary Value Problems (IBVPs). Specifically, the author aims to develop a class of FD methods called "Error Inhibiting Schemes (EIS)" such that the convergence order of the global error is higher than that of the truncation error.
### Problem Background
When traditional FD methods are used to solve Partial Differential Equations (PDEs), their global errors are usually of the same order as the truncation errors. For example, in the classical FD scheme, if the truncation error is \(O(h^p)\), then the global error is also \(O(h^p)\). However, for some applications, it is hoped that the global error can converge more quickly, that is, the order of the global error is higher than that of the truncation error.
### Research Motivation
The author was inspired by Nodal - basis Discontinuous Galerkin (DG) methods. These methods can be regarded as Block Finite Difference (BFD) schemes and exhibit supra - convergence in some cases. Although the truncation error may be low, the overall scheme can still provide higher accuracy. Therefore, the author's goal is to design a BFD scheme such that the convergence order of the global error is higher than that of the truncation error.
### Main Contributions
1. **Generalization to IBVPs**: The main contribution of this paper lies in considering IBVPs with Dirichlet or Neumann boundary conditions. For such problems, the main challenge in designing high - order schemes is the design of the boundary stencil. The author has successfully solved this problem by introducing an error - inhibiting mechanism.
2. **Error Analysis and Stability Proof**: The author provides a detailed error analysis, including proofs of stability and convergence. They show how to use the algebraic structure to separate the solution space and the truncation error space, thereby constructing a numerical mechanism that inhibits the accumulation of truncation errors.
3. **Numerical Verification**: Through multiple numerical experiments, the effectiveness of the proposed method is verified. The experimental results show that the new method is superior to traditional FD methods in terms of global error convergence.
### Specific Implementation
- **Third - order Scheme under Periodic Boundary Conditions**: The author first introduces a third - order BFD scheme for periodic boundary conditions and gives a complete error analysis.
- **Third - order and Fifth - order Schemes under Non - periodic Boundary Conditions**: Further, this method is generalized to non - periodic boundary conditions (such as Dirichlet and Neumann boundary conditions), and third - order and fifth - order BFD schemes are designed respectively.
- **Boundary Treatment**: Extrapolation is used near the boundary to generate ghost points to ensure the accuracy and stability of the boundary conditions.
### Conclusion
Through the above research, the author has successfully designed a class of error - inhibiting schemes (EIS) such that the convergence order of the global error is higher than that of the truncation error. This result not only improves the accuracy of the numerical solution but also provides new ideas and methods for solving more complex IBVPs.