monotonescheme 0.0.17

Python wrapper for C++ codes for the monotone scheme for curvature-driven PDEs

Monotone schemes for curvature-driven PDEs

by Jeff Calder (UMN) and Wonjun Lee (UMN)

• Paper: arXiv
• Jeff Calder, School of Mathematics, University of Minnesota: website
• Wonjun Lee, Institute for Mathematics and Its Applications, Uniersity of Minnesota: website

---

Introduction

This repository contains c++ and python codes for running the monotone algorithm to solve curvature-driven PDEs. Here are list of PDEs that can be solved using this algorithm. Let $\Omega = [0,1]^d$ be a domain in $\mathbb{R}^d$ and $\partial \Omega$ be a boundary of $\Omega$.

Eikonal equation

$$ \begin{align*} |\nabla u(x)| &= f(x), && x \in \Omega \ x &= 0, && x \in \partial \Omega \end{align*} $$

Mean curvature PDE

$$ \begin{align*} |\nabla u(x)|\kappa(x) &= f(x), && x \in \Omega \ x &= 0, && x \in \partial \Omega \end{align*} $$ where $\kappa(x) = - \text{div}\left( \frac{\nabla u}{|\nabla u|} \right)$ is the mean curvature of the level set surface of $u$ passing through $x$.

Affine flows PDE

$$ \begin{align*} |\nabla u(x)|\kappa(x)+^{\alpha} &= f(x), && x \in \Omega \ x &= 0, && x \in \partial \Omega \end{align*} $$ where $\alpha \in (0,1]$ is a constant depending on the dimension $d$ and $(t)+ := \max(0,t)$.

Tukey Depth

$$ \begin{align*} |\nabla u(x)| &= \int_{(y-x)\cdot \nabla u(x) = 0} \rho(y) dS(y), && x \in \Omega \ x &= 0, && x \in \partial \Omega \end{align*} $$

---

Tutorial

Prerequisites

• pip
• python >= 3.6

Follow this link to see the instruction for the installation of pip: https://pip.pypa.io/en/stable/installation/.

Installing the package

Install the package by running the following command:

    pip install monotonescheme

Running the codes

You can find the example python script files and notebook files in tests folder. The notebook files in the folder solves the following problems:

1. Affine flows in 2D Cartesian grid.

• tests/affine_PDE_2D.ipynb
• tests/affine_PDE_2D.py

2. Tukey depth eikonal equation in 2D Cartesian grid.

• tests/tukey_PDE_2D.ipynb
• tests/tukey_PDE_2D.py

3. Motion by curvature PDE in 3D Cartesian grid.

• tests/curvature_PDE_3D.ipynb
• tests/curvature_PDE_3D.py

4. Eikonal equation and Tukey depth eikonal equation in unstructure grids

• tests/Eikonal_PDE_graph.ipynb
• tests/tukey_PDE_graph.ipynb
• tests/Eikonal_PDE_graph.py
• tests/tukey_PDE_graph.py