OrdinaryDiffEq.jl icon indicating copy to clipboard operation
OrdinaryDiffEq.jl copied to clipboard

Numerov algorithm

Open feanor12 opened this issue 7 years ago • 6 comments

Is there a implemented algorithm which is equal to Numerov's algorithm ?

Where should i look to implement it?

feanor12 avatar Oct 04 '18 07:10 feanor12

There currently is no implementation of it. It would go into OrdinaryDiffEq.jl.

If you'd like to give it a shot, the devdocs could be helpful: http://devdocs.juliadiffeq.org/latest/contributing/adding_algorithms.html . Feel free to ask in our chatroom as well. The symplectic integrator codes would be a good place to start at. For the implicit equations you might want to look at how ImplicitEuler is done.

@YingboMa Xref: https://github.com/JuliaDiffEq/OrdinaryDiffEq.jl/issues/331

ChrisRackauckas avatar Oct 04 '18 11:10 ChrisRackauckas

I tried to build something, but I few things don't work. How to I get different values for u,uprev,uprev2 etc.

Here is the code I wrote until now: https://gist.github.com/feanor12/867f0b04a732824f3bff6846899bb167

feanor12 avatar Oct 17 '18 14:10 feanor12

I updated the code again rev 4 should work. https://gist.github.com/feanor12/867f0b04a732824f3bff6846899bb167

feanor12 avatar Oct 17 '18 20:10 feanor12

The linear version would need to make use of linear operator tools which aren't very well documented right now. That one probably needs some more interface extensions to work. @YingboMa might want to look into it. But I would start with the nonlinear version. That should be a not too difficult and would just utilize a lot of the structure from the nonlinear solvers of the first order ODEs

ChrisRackauckas avatar Oct 26 '18 00:10 ChrisRackauckas

To what extent is this issue still "open" now? @feanor12 has a code that they say works. @ChrisRackauckas shall it go in OrdinaryDiffEq.jl?

ndattani avatar Feb 07 '23 22:02 ndattani

I think it's still wide open.

ChrisRackauckas avatar Feb 15 '23 06:02 ChrisRackauckas