Re: math 128A programming set 3

10 views
Skip to first unread message

Soroosh Yazdani

unread,
Oct 27, 2006, 11:43:50 AM10/27/06
to Nicole Aolin Li, math128a...@groups.l.google.com
If you don't implement Adaptive Quadrature in problem one, then you
can use the code provided on the book website.
You should implement everything else yourself.

Soroosh
On Fri, Oct 27, 2006 at 12:00:01AM -0700, Nicole Aolin Li wrote:
> Hi Soroosh,
>
> For the programming assignment, are we allowed to use the matlab code
> provided on the book website? Or we have to write it by ourselves even
> though there are those codes?
>
> Thanks,
>
> Aolin Li

Soroosh Yazdani

unread,
Oct 31, 2006, 7:18:28 PM10/31/06
to Nicole Aolin Li, math128a...@groups.l.google.com
Take the results for adaptive quadrature to be correct to the accuracy
that you ask for, since that's gauranteed. This way you can have something
that is pretty close to exact if you choose your tolerance small enough
(say 1e-10). You can check how long this takes as a measure of efficiency.

The other methods are really efficient, so just measure the accuracy by
comparing the values with the values you get from part (a).
If one of them gives you good answers for a while, and then gives you nonsense
answers, then that is unstable.

Soroosh


On Tue, Oct 31, 2006 at 02:25:59AM -0800, Nicole Aolin Li wrote:
> Hi Soroosh,
>

> For Problem 2(d), I'm wondering how we're supposed to compare the three
> methods with respect to accuracy. Are there exact values we can use to
> serve as standards?
>
> Thanks,
>
> Aolin Li

Reply all
Reply to author
Forward
0 new messages