Re: [boost] Fwd: Regarding GSoc'14 Boost.odeint idea

Subject: Re: [boost] Fwd: Regarding GSoc'14 Boost.odeint idea
From: Karsten Ahnert (karsten.ahnert_at_[hidden])
Date: 2014-03-10 17:17:12

> BDF implicit routine is of prime importance when it comes to stiff
> systems, but are we expected to test its flexibility on other systems
> as well? Or is it that other implicit methods shall be implemented for
> other system.  Eg. Implicit Euler method or maybe the Adams-Moulton
> method ?

One main use case for implicit methods in general are stiff systems.
Therefore, I think one should focus on stiff systems. Of course, it is
easy to test it with non-stiff ODEs. An implicit Euler method already
exist. But it is limited to Boost.ublas for the state and matrix type at
the moment. One goals is to allow for arbitrary state types and the
Implicit Euler is surely the most simple method to test if the state
type can simply be exchanged.