A limited memory BFGS solver for phase-field fracture simulations: code gallery or tutorial step?

[Tao Jin]

Dear deal.ii developers and community,

Recently, our group developed a limited-memory BFGS solver (a type of quasi Newton method) as a monolithic approach to solve the phase-field crack propagation problems via deal.ii. My question is, which is a better way to share the code with the deal.ii community? Should we contribute the code in the deal.ii code-gallery? Or could the code be potentially included as a tutorial step?

Here is more information:

Our paper can be found here: https://onlinelibrary.wiley.com/doi/full/10.1002/nme.7572.

And our source code and input files for all the numerical examples are hosted here:

https://github.com/taojinllnl/Phasefield_L-BFGS_monolithic_solver

The unique features of our code comparing with existing tutorial steps:

1. The implementation of the limited-memory BFGS solver (a type of quasi Newton solver) that can be used in various non-convex optimization problems.

2. The limited-memory BFGS solver is combined with an adaptive refinement technique, which involves solution transfer and projection from coarse meshes to refined meshes.

3. A widely adopted phase-field model with various 2D and 3D benchmark problems.

Thanks for your feedback.

Tao Jin, PhD

University of Ottawa

[Luca Heltai]

Dear Tao,

I personally think this would be a perfect fit for our code gallery!

Thanks for sharing.

L.

> --
> The deal.II project is located at http://www.dealii.org/
> For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en
> ---
> You received this message because you are subscribed to the Google Groups "deal.II User Group" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to dealii+un...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/dealii/7382c669-b61e-478f-adcc-d192657dece3n%40googlegroups.com.