Gianluca Frasca-Caccia,Peter E. Hydon
Abstract:There are several well-established approaches to constructing finite difference schemes that preserve global invariants of a given partial differential equation. However, few of these methods preserve more than one conservation law locally. A recently-introduced strategy uses symbolic algebra to construct finite difference schemes that preserve several local conservation laws of a given scalar PDE in Kovalevskaya form. In this paper, we adapt the new strategy to PDEs that are not in Kovalevskaya form and to systems of PDEs. The Benjamin-Bona-Mahony equation and a system equivalent to the nonlinear Schroedinger equation are used as benchmarks, showing that the strategy yields conservative schemes which are robust and highly accurate compared to others in the literature.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to develop finite - difference schemes for partial differential equations (PDEs) that can simultaneously maintain multiple local conservation laws. Specifically, the author aims to:
1. **Expand existing methods**: Most of the existing methods can only maintain one or a few global invariants, and few methods can maintain multiple conservation laws at the local level. The goal of this paper is to construct finite - difference schemes that can maintain multiple local conservation laws by introducing new strategies.
2. **Apply to PDEs and PDE systems not in Kovalevskaya form**: Previous methods were mainly aimed at PDEs in Kovalevskaya form. This paper attempts to generalize these methods to PDEs and PDE systems not in Kovalevskaya form.
3. **Improve numerical accuracy and robustness**: By constructing conservative schemes, ensure that the numerical solutions maintain high accuracy and stability in long - time simulations. In particular, for the applications of the Benjamin - Bona - Mahony (BBM) equation and two real PDE systems equivalent to the nonlinear Schrödinger (NLS) equation, the effectiveness and superiority of the new method are demonstrated.
### Specific problem description
- **Maintaining multiple local conservation laws**: Many physical phenomena are described by partial differential equations, which usually have multiple conservation laws (such as mass, momentum, and energy conservation). Traditional numerical methods often can only maintain global invariants and cannot well preserve local conservation laws. This paper proposes a new strategy, using symbolic algebraic techniques to construct finite - difference schemes that can simultaneously maintain multiple local conservation laws.
- **Applicable to a wider range of PDE types**: In addition to the common PDEs in Kovalevskaya form, this paper also explores how to apply this method to PDEs and PDE systems not in Kovalevskaya form. This greatly broadens the application range of this method.
- **Verification and comparison**: By using the BBM equation and the NLS equation as benchmark test cases, it is proved that the conservative schemes generated by the proposed strategy are numerically more accurate and robust than existing methods.
### Mathematical formula representation
In order to clearly express the mathematical concepts involved, the following are some key formulas mentioned in the paper:
- **Local conservation law**:
\[
\text{Div} F = D_x\{F(x,t,[u])\} + D_t\{G(x,t,[u])\}
\]
where \(D_x\) and \(D_t\) are the total derivative operators with respect to \(x\) and \(t\) respectively.
- **Discrete conservation law**:
\[
\text{Div} \tilde{F} = D_m\{\tilde{F}(m,n, [u])\} + D_n\{\tilde{G}(m,n, [u])\}
\]
- **Characteristic form**:
\[
\text{Div} F = A Q
\]
where \(Q\) is the characteristic vector.
Through these formulas, the author constructs finite - difference schemes that can maintain multiple local conservation laws and verifies their effectiveness and superiority through numerical experiments.