Question about Modified Newton-Raphson and UMFpack solver

For developers writing C++, Fortran, Java, code who have questions or comments to make.

Moderators: silvia, selimgunay, Moderators

Post Reply
TsarpalisD
Posts: 8
Joined: Tue Oct 18, 2016 9:06 am
Location: National Technical University of Athens

Question about Modified Newton-Raphson and UMFpack solver

Post by TsarpalisD »

Hello Osees community,
I have a question about the implementation of Modified Newton-Raphson algorithm (or equivalently Newton with -initial) in Osees. To be more specific in the method "int ModifiedNewton::solveCurrentStep(void)" the method theSOE->solve() is invoked which performs umfpack_di_numeric (i.e. LU decomposition) and then umfpack_di_solve (i.e. solving the system). To my understanding the purpose of mNR is do avoid the expensive LU decomposition, so isn't it an overkill to perform it in each step?
Thanks in advance,
Tsarpalis Dimitris
Post Reply