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

-- 
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 [email protected].
To view this discussion visit 
https://groups.google.com/d/msgid/dealii/7382c669-b61e-478f-adcc-d192657dece3n%40googlegroups.com.

Reply via email to