基于卷积神经网络模型数值求解双曲型偏微分方程的研究

高普阳; 赵子桐; 杨扬

doi:10.21656/1000-0887.420050

基于卷积神经网络模型数值求解双曲型偏微分方程的研究

doi: 10.21656/1000-0887.420050

1. 长安大学理学院，西安 710064;
2. 西北工业大学航空学院，西安 710072

基金项目:

国家自然科学基金(11901051

陕西省自然科学基础研究计划青年项目(2020JQ-338；2019JQ-625)

11971075)

详细信息

作者简介:
高普阳(1991—)，男，讲师，博士（通讯作者. E-mail: gaopuyang@chd.edu.cn）.

通讯作者:
高普阳(1991—)，男，讲师，博士（通讯作者. E-mail: gaopuyang@chd.edu.cn）.

中图分类号: O29
计量
- 文章访问数: 729
- HTML全文浏览量: 122
- PDF下载量: 146
- 被引次数: 0
出版历程
- 收稿日期: 2021-02-23
- 修回日期: 2021-04-21
- 网络出版日期: 2021-09-29

Study on Numerical Solutions to Hyperbolic Partial Differential Equations Based on the Convolutional Neural Network Model

1. School of Sciences, Chang’an University, Xi’an 710064, P.R.China;
2. School of Aeronautics, Northwestern Polytechnical University, Xi’an 710072, P.R.China

Funds:

The National Natural Science Foundation of China(11901051

11971075)

摘要

摘要: 人工神经网络近年来得到了快速发展，将此方法应用于数值求解偏微分方程是学者们关注的热点问题.相比于传统方法其具有应用范围广泛（即同一种模型可用于求解多种类型方程）、网格剖分条件要求低等优势，并且能够利用训练好的模型直接计算区域中任意点的数值.该文基于卷积神经网络模型，对传统有限体积法格式中的权重系数进行优化，以得到在粗粒度网格下具有较高精度的新数值格式，从而更适用于复杂问题的求解.该网络模型可以准确、有效地求解Burgers方程和level set方程，数值结果稳定，且具有较高数值精度.
- 卷积神经网络模型 /
- Burgers方程 /
- level set方程 /
- 有限体积法
Abstract: In recent years, artificial neural networks developed rapidly. Application of this method to partial differential equations became a new idea for exploring numerical solutions to differential equations. Compared with the traditional methods, it has some advantages, such as a wide range of applications (i.e. the same model can be used to solve multiple types of equations) and low meshing requirements. In addition, the trained model can be directly used to calculate the numerical solution at any point in the computation domain. The weight coefficients in the traditional finite volume method were optimized based on the convolutional neural network model to get a new numerical scheme with highresolution results on the coarse grid. The proposed model helps solve the Burgers and level set equations efficiently and stably with high accuracy.
- convolutional neural network model /
- Burgers equation /
- level set equation /
- finite volume method

HTML全文

参考文献(19)

[2]RUDY S, ALLA A, BRUNTON S L, et al. Data-driven identification of parametric partial differential equations[J].SIAM Journal on Applied Dynamical Systems,2019,18(2): 643-660.

LI Y, DUAN J, LIU X. A machine learning framework for computing the most probable paths of stochastic dynamical systems[J].Physical Review E,2021,103(1): 012124.

[3]BECK C, WEINAN E, JENTZEN A. Machine learning approximation algorithms for high-dimensional fully nonlinear partial differential equations and second-order backward stochastic differential equations[J].Journal of Nonlinear Science,2019,29(4): 1563-1619.

[4]DONG H, NIE Y, CUI J, et al. A wavelet-based learning approach assisted multiscale analysis for estimating the effective thermal conductivities of particulate composites[J].Computer Methods in Applied Mechanics and Engineering,2021,374: 113591.

[5]KUMAR M, YADAV N. Multilayer perceptrons and radial basis function neural network methods for the solution of differential equations: a survey[J].Computers & Mathematics With Applications,2011,62(10): 3796-3811.

[6]YAZDI H S, PAKDAMAN M, MODAGHEGH H. Unsupervised kernel least mean square algorithm for solving ordinary differential equations[J].Neurocomputing,2011,74(12/13): 2062-2071.

[7]MISHRA S. A machine learning framework for data driven acceleration of computations of differential equations[J].Mathematics in Engineering,2018,1(11): 118-146.

[8]PISCOPO M L, SPANNOWSKY M, WAITE P. Solving differential equations with neural networks: applications to the calculation of cosmological phase transitions[J].Physical Review D,2019,100(1): 016002.

[9]WINOVICH N, RAMANI K, LIN G. ConvPDE-UQ: convolutional neural networks with quantified uncertainty for heterogeneous elliptic partial differential equations on varied domains[J].Journal of Computational Physics,2019,394: 263-279.

[10]GUINOT V.Godunov-Type Schemes: an Introduction for Engineers[M]. Elsevier, 2003.

[11]GARDINER T A, STONE J M. An unsplit Godunov method for ideal MHD via constrained transport[J].Journal of Computational Physics,2005,205(22): 509-539.

[12]VAN LEER B. Towards the ultimate conservative difference scheme Ⅴ: a second-order sequel to Godunov’s method[J].Journal of Computational Physics,1979,32(1): 101-136.

[13]RUF A M, SANDE E, SOLEM S. The optimal convergence rate of monotone schemes for conservation laws in the Wasserstein distance[J].Journal of Scientific Computing,2019,80:1764-1776.

[14]毕卉. 基于Godunov格式的求解Burgers方程的有限差分法[D]. 硕士学位论文. 哈尔滨: 哈尔滨工业大学, 2006.(BI Hui. Finite difference method for solving Burgers equation based on Godunov scheme[D]. Master Thesis. Harbin: Harbin Institute of Technology, 2006.(in Chinese))

[15]KINGMA D P, BA J L. Adam: a method for stochastic optimization[R/OL]. 2014.[2021-03-23]. https://perso.ensta-paris.fr/~pcarpent/StochOpt/PDF/Articles/Kingma_arXiv_2017.pdf.

[16]BAR-SINAI Y, HOYER S, HICKEY J, et al. Learning data-driven discretizations for partial differential equations[J].Proceedings of the National Academy of Ences,2019,116(31):15344-15349.

[17]包立平, 洪文珍. 一维弱噪声随机Burgers方程的奇摄动解[J]. 应用数学和力学， 2018,39(1): 113-122.(BAO Liping, HONG Wenzhen. Singular perturbation solutions to 1D stochastic burgers equations under weak noises[J].Applied Mathematics and Mechanics,2018,39(1): 113-122.(in Chinese))

[18]LIAO W. An implicit fourth-order compact finite difference scheme for one-dimensional Burgers’equation[J].Applied Mathematics and Computation,2008,206(2): 755-764.

[19]MARCHANDISE E, REMACLE J F, CHEVAUGEON N. A quadrature-free discontinuous Galerkin method for the level set equation[J].Journal of Computational Physics,2006,212(1): 338-357.

施引文献

资源附件(0)

访问统计