Skip to content

Multi-scale DNN for solving elliptic multi-scale PDEs by shifting the input data in narrow- range into large-range, then fed the transformed data into the DNN pipline

License

Notifications You must be signed in to change notification settings

xuliang5115/MscaleDNN_torch

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MscaleDNN_torch

Solving a class of elliptic multi-scale PDEs by shifting the input data in narrow-range into large-range, then fed the transformed data into the DNN pipline

Requirement

python = 3.9

pytorh =1.12.1

cuda = 11.6

Corresponding Papers

A multi-scale DNN algorithm for nonlinear elliptic equations with multiple scales

created by Xi-An Li, Zhi-Qin John, Xu and Lei Zhang

[Paper]

Ideas

This work exploited the technique of shifting the input data in narrow-range into large-range, then fed the transformed data into the DNN pipline.

Abstract:

Algorithms based on deep neural networks (DNNs) have attracted increasing attention from the scientific computing community. DNN based algorithms are easy to implement, natural for nonlinear problems, and have shown great potential to overcome the curse of dimensionality. In this work, we utilize the multi-scale DNN-based algorithm (MscaleDNN) proposed by Liu, Cai and Xu (2020) to solve multi-scale elliptic problems with possible nonlinearity, for example, the p-Laplacian problem. We improve the MscaleDNN algorithm by a smooth and localized activation function. Several numerical examples of multi-scale elliptic problems with separable or non-separable scales in low-dimensional and high-dimensional Euclidean spaces are used to demonstrate the effectiveness and accuracy of the MscaleDNN numerical scheme.

Noting:

The original codes are found in https://github.com/Blue-Giant/MscaleDNN_tf1, and the coorresponding dada are provided in this url.

The matlab codes in 2D辅助matlab代码/p=2 are useful for E1,E2,E3 and E4.

The matlab codes in 2D辅助matlab代码/p=3Forier_scale are useful for E5.

The matlab codes in 2D辅助matlab代码/p=3Subspace are useful for E6.

About

Multi-scale DNN for solving elliptic multi-scale PDEs by shifting the input data in narrow- range into large-range, then fed the transformed data into the DNN pipline

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 93.8%
  • MATLAB 6.2%