MathGroup Archive 2007

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Precision when solving a system of differential equations

  • To: mathgroup at smc.vnet.net
  • Subject: [mg81841] Re: Precision when solving a system of differential equations
  • From: Jens-Peer Kuska <kuska at informatik.uni-leipzig.de>
  • Date: Thu, 4 Oct 2007 04:34:35 -0400 (EDT)
  • Organization: Uni Leipzig
  • References: <fdve27$t1e$1@smc.vnet.net>
  • Reply-to: kuska at informatik.uni-leipzig.de

Hi,

you can't because NDSolve[] determine the step size from
the precision you set.
You can reduce the precision that NDSolve[] use but that
is usual not a good idea.

In a stiff equation one has typical to made some very small steps
but when the system relaxes the step size can grow. So
it is not a good idea to set a lower limit for the step
size. You should trust the step size control.

Regards
   Jens


Marc VIAL wrote:
> Hi,
> 
> I have a system with three differential equations with time as my variable.
>  I would like to control the increment time step as some of my function are
>  very stiff. Ideally I would like to use between 10 and 20 millions steps. 
> How can I do that?
> (version used Mathematica 6.0.1)
> 
> Thanks for help
> Marc
> _________________________________________________________________
> 


  • Prev by Date: Interpolation does not quite interpolate
  • Next by Date: Re: BarChart - intensity of colour increases up the bar
  • Previous by thread: Precision when solving a system of differential equations
  • Next by thread: Delayed Derivative Operator