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

Re: Inverse, numerical precision, simulation dynamical system

• To: mathgroup at smc.vnet.net
• Subject: [mg77313] Re: [mg77211] Inverse, numerical precision, simulation dynamical system
• From: DrMajorBob <drmajorbob at bigfoot.com>
• Date: Wed, 6 Jun 2007 07:28:59 -0400 (EDT)
• References: <5452490.1181096373813.JavaMail.root@m35>
• Reply-to: drmajorbob at bigfoot.com

We haven't seen any of your code, so how could we know where to improve it?

Certainly N[Inverse[x],1000] can't be faster than Inverse[x], since both  
require computing the Inverse.

Bobby

On Tue, 05 Jun 2007 06:01:17 -0500, kristoph <kristophs.post at web.de> wrote:

> Dear Experts,
> I am quiet new at the subject and would be grateful for a hint.
>
> I'm simulating a stochastic dynamical system which includes
> calculating the inverse of (for now) a 2-by-2 matrix. At each
> iteration these values are inputs for the next. Simply using
> Inverse[.] does a very poor job considering timing.
>
> To speed up the process I used N[Inverse[.],1000]. It is better but
> sometimes even a precision of 1000 is not enough resulting in an
> "significant numerical error".
> The dynamical system consists of 1000 iterations and I'm simulating it
> 500-800 times for statistical reasons thus, time is an issue.
>
> Instead of using N[Inverse[.],1250] and so on for all iterations it
> would be great if there is a way to speed up time and avoiding the
> error.
>
> Thank you very much and all the best.
> Kristoph
>
>
>



-- 
DrMajorBob at bigfoot.com