Skip to content

A setup for solving shape optimization problems depending on a partial differential equation using boundary integral methods and the Python library JAX. The code was developed as part of a Bachelor thesis project in Engineering Mathematics at the Royal Institute of Technology in Spring 2024 by Rebecka Johansson and Asta Stensson.

Notifications You must be signed in to change notification settings

astastensson/PDE-constrained-Shape-Optimization-with-Boundary-Integral-Methods

Repository files navigation

PDE-constrained Shape Optimization with Boundary Integral Methods

A setup for solving shape optimization problems depending on a partial differential equation using boundary integral methods and the Python library JAX. The PDE in question is Laplaces equation with a Dirichlet boundary condition. Any maximization problem where the goal function M depends on the solution u of Laplaces equations can be put into the setup. The PDE is then solved using boundary integral methods, and M is optimized with gradient descent using automatic differentiation in JAX. Some experiments and examples can be found in the Jupyter Notebooks. The code was developed as part of a Bachelor thesis project by Rebecka Johansson and Asta Stensson for the Engineering Mathematics program at The Royal Institute of Technology (KTH) in spring 2024.

PDE-villkorad formoptimering med randintegralmetoder

Ett ramverk för att slösa formoptimeringsproblem som beror på en partiell differentialekvation med randintegralmetoder och Pythonbiblioteket JAX. PDE:n i fråga är Laplaces ekvation med Dirichletvillkor. Vilket maximeringsproblem som helst där målfunktionen M beror på lösningen u av Laplaces ekvation kan undersökas. PDE:n är löst med randintegralmetoder och M optimeras sedan med gradientstegning med hjälp av JAX automatiska differentiering. Några experiment och exempel kan hittas i Jupyter Notebooksen. Koden är utvecklad som ett kandidatexamensarbete av Rebecka Johansson och Asta Stensson för civilingenjörsprogrammet i teknisk matematik på Kungliga tekniska högskolan (KTH) under våren 2024.

About

A setup for solving shape optimization problems depending on a partial differential equation using boundary integral methods and the Python library JAX. The code was developed as part of a Bachelor thesis project in Engineering Mathematics at the Royal Institute of Technology in Spring 2024 by Rebecka Johansson and Asta Stensson.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published