https://github.com/milescb/solve_pdes_with_pinn
Numerical solutions of several PDEs using Physics-Informed Neural Networks
https://github.com/milescb/solve_pdes_with_pinn
ml pde physics-informed-neural-networks sci-ml
Last synced: over 1 year ago
JSON representation
Numerical solutions of several PDEs using Physics-Informed Neural Networks
- Host: GitHub
- URL: https://github.com/milescb/solve_pdes_with_pinn
- Owner: milescb
- License: mit
- Created: 2022-10-07T08:42:52.000Z (almost 4 years ago)
- Default Branch: master
- Last Pushed: 2023-08-10T08:12:05.000Z (almost 3 years ago)
- Last Synced: 2023-08-10T10:29:02.008Z (almost 3 years ago)
- Topics: ml, pde, physics-informed-neural-networks, sci-ml
- Language: Julia
- Homepage:
- Size: 1.91 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Solving PDEs with Physics-Informed Neural Networks
The following repository contains examples using Physics-Informed Neural Networks (PINN) to solve PDEs. We use the package `NeuralPDE.jl` to solve. This work is part of my senior capstone for Lawrence University in Appleton, WI. I have documented the techniques used here along with appropriate background information [in this paper](https://github.com/lvb5/senior_project_paper/blob/master/main.pdf).
## List of examples
Find a list of the example problems we have solved or are working on below
- Integral-PDE
- PDAE
- Linear, homogeneous PDE
- Einstein field equations to find Schwarzschild metric. This problem has been solved in its most simple case. Continued work is needed in extending the problem. Find more information in the README in `src/solve_einstein`.
## Running the Code
Several required packages are included in the `Project.toml` to allow one to run this code out of the box. You can use the environment in this repo to quickly load the correct versions of the packages by running
```
julia> using Pkg
julia> Pkg.instantiate()
```
Then, to run the code either activate the environment and run from the REPL, or run the scripts with
```
julia --project .jl
```
from terminal. Note that `Julia` needs to be in your path for this to work.
## Contribution
Pull requests are encouraged!