search

SciML / NonlinearSolve.jl

High-performance and differentiation-enabled nonlinear solvers (Newton methods), bracketed rootfinding (bisection, Falsi), with sparsity and Newton-Krylov support.
https://docs.sciml.ai/NonlinearSolve/stable/
MIT License
221 stars 39 forks source link

FixedPointSolvers: Recommendations and Wrappers #337

Closed avik-pal closed 7 months ago

avik-pal commented 7 months ago

Fixes #76, Fixes #78

codecov[bot] commented 7 months ago

Codecov Report

Attention: 11 lines in your changes are missing coverage. Please review.

Comparison is base (d723578) 88.79% compared to head (29917d8) 88.47%.

Files Patch % Lines
src/extension_algs.jl 73.07% 7 Missing :warning:
ext/NonlinearSolveSpeedMappingExt.jl 83.33% 3 Missing :warning:
ext/NonlinearSolveFixedPointAccelerationExt.jl 96.42% 1 Missing :warning:
Additional details and impacted files ```diff @@ Coverage Diff @@ ## master #337 +/- ## ========================================== - Coverage 88.79% 88.47% -0.32% ========================================== Files 24 26 +2 Lines 1972 2048 +76 ========================================== + Hits 1751 1812 +61 - Misses 221 236 +15 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

avik-pal commented 7 months ago

@ChrisRackauckas what do you think about introducing a FixedPointProblem type? I don't quite like the idea of doing the back and forth of adding and subtracting u, also I feel it is easy to make a mistake while providing the custom jacobian and such

ChrisRackauckas commented 7 months ago

@ChrisRackauckas what do you think about introducing a FixedPointProblem type? I don't quite like the idea of doing the back and forth of adding and subtracting u, also I feel it is easy to make a mistake while providing the custom jacobian and such

Yeah that sounds like a good thing to add at some point, where some methods would work on that natively and others would convert.