Abstract:In this paper, we introduce the Deep Finite Volume Method (DFVM), an innovative deep learning framework tailored for solving high-order (order \(\geq 2\)) partial differential equations (PDEs). Our approach centers on a novel loss function crafted from local conservation laws derived from the original PDE, distinguishing DFVM from traditional deep learning methods. By formulating DFVM in the weak form of the PDE rather than the strong form, we enhance accuracy, particularly beneficial for PDEs with less smooth solutions compared to strong-form-based methods like Physics-Informed Neural Networks (PINNs). A key technique of DFVM lies in its transformation of all second-order or higher derivatives of neural networks into first-order derivatives which can be comupted directly using Automatic Differentiation (AD). This adaptation significantly reduces computational overhead, particularly advantageous for solving high-dimensional PDEs. Numerical experiments demonstrate that DFVM achieves equal or superior solution accuracy compared to existing deep learning methods such as PINN, Deep Ritz Method (DRM), and Weak Adversarial Networks (WAN), while drastically reducing computational costs. Notably, for PDEs with nonsmooth solutions, DFVM yields approximate solutions with relative errors up to two orders of magnitude lower than those obtained by PINN. The implementation of DFVM is available on GitHub at \href{<a class="link-external link-https" href="https://github.com/Sysuzqs/DFVM" rel="external noopener nofollow">this https URL</a>}{<a class="link-external link-https" href="https://github.com/Sysuzqs/DFVM" rel="external noopener nofollow">this https URL</a>}.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the challenges encountered by traditional numerical methods and existing deep - learning methods when solving partial differential equations (PDEs) in the case of high - dimensional and non - smooth solutions. Specifically:
1. **High - dimensional problems**: Traditional numerical methods such as the finite - difference method, the finite - element method, and the finite - volume method face the "curse of dimensionality" problem when dealing with high - dimensional PDEs. That is, as the dimension increases, the number of unknowns grows exponentially, resulting in a sharp increase in computational cost.
2. **Non - smooth solutions**: The solutions of many practical physical problems may not be smooth enough to strictly satisfy the strong form of PDEs, but can satisfy the weak form of PDEs. Existing deep - learning methods based on strong - form loss functions (such as Physics - Informed Neural Networks, PINNs) perform poorly when dealing with such problems because they rely on the smoothness of the solutions.
To solve these problems, the author proposes the Deep Finite Volume Method (DFVM), a new deep - learning framework specifically designed for solving high - order (order ≥ 2) PDEs. The main innovations of DFVM include:
- **Weak - form loss function**: DFVM designs a new loss function based on the weak form of PDEs, using local conservation laws to construct the loss function. This makes DFVM more accurate when dealing with non - smooth solutions. In particular, when the solution is not smooth enough, the relative error of DFVM can be two orders of magnitude lower than that of PINNs based on the strong form.
- **High - efficiency computation**: DFVM significantly reduces computational overhead by converting all second - order or higher - order derivatives of the neural network into first - order derivatives. This method is especially suitable for solving high - dimensional PDEs because first - order derivatives can be directly calculated by automatic differentiation (AD), while the computational complexity of second - order derivatives increases significantly with the increase of dimension.
- **Wide applicability**: DFVM can be applied not only to second - order PDEs, but also can be extended to higher - order PDEs by introducing intermediate variables. In addition, DFVM performs well in dealing with asymmetric PDEs, high - order PDEs, and high - dimensional PDEs, making it a general - purpose tool for solving complex PDEs.
In summary, by proposing DFVM, this paper aims to provide an efficient, accurate, and widely applicable deep - learning method to address the deficiencies of traditional numerical methods and existing deep - learning methods in dealing with high - dimensional and non - smooth - solution PDEs.