search

SciML / SimpleNonlinearSolve.jl

Fast and simple nonlinear solvers for the SciML common interface. Newton, Broyden, Bisection, Falsi, and more rootfinders on a standard interface.
https://docs.sciml.ai/NonlinearSolve/stable/
MIT License
68 stars 26 forks source link

ITP default values differ from Tuning Parameters #108

Closed judober closed 3 months ago

judober commented 11 months ago

First of all, thank you for this great package!

When looking through the documentation of the ITP-Method (and the code), I noticed that the default values of the parameters are different between the header (==code) and the 'Tuning Parameters' section which seems to be wrong. Is there a reason that k2 is not an int as recommended in the text?

ChrisRackauckas commented 11 months ago

I think the docstrings and such just didn't get updated. On the int thing, maybe @DaniGlez can chime in

DaniGlez commented 11 months ago

This did indeed pop up in the discussion we had with Oscar, I think we just did not act on this specific point (probably because I was not sure if the default values were coming from some analysis I was not aware of): https://github.com/SciML/DiffEqBase.jl/pull/917#issuecomment-1646235996

But it seems indeed sensible to set it to 2, and we can port that to the InternalITP from DiffEqBase as well