Re: Numerical Integration
- From: "John D'Errico" <woodchips@xxxxxxxxxxxxxxxx>
- Date: Tue, 22 Jul 2008 15:18:04 +0000 (UTC)
"selim selim" <fthsel@xxxxxxxxxxx> wrote in message
<g64se3$15d$1@xxxxxxxxxxxxxxxxxx>...
Hallo,
I have a rectangular domain and equally spaced points on
this domain. At each point I have some variables
(velocity , temperature etc..)
I want to evaluate integral
in the form integral (T(x,y)*u(x,y)+u(x,y)^2) over domain.
I dont have any functions for temperature and velocity ,
just discrete data on the equally spaced points.
How could I do that integral accuretly?
Summation could be , but in order to be more precise what
can I do? I need matlab commands, or a code? (Chebyschev,
Gauss Legendr. etc. could be?)
If you only have fixed points, then forget about
something like Gauss-Legendre, which will
specify the sample points.
You can use any Newton-Cotes rule, of which
trapezoidal and Simpson's rule are good
examples. trapz does trapezoidal rule, and
you can find several Simpson's rule variants
on the file exchange.
John
.
- Follow-Ups:
- Re: Numerical Integration
- From: selim selim
- Re: Numerical Integration
- References:
- Numerical Integration
- From: selim selim
- Numerical Integration
- Prev by Date: Re: Saving a Full running GUI figure
- Next by Date: Re: vectorized computation in C++ such as those in Matlab (Matlab to
- Previous by thread: Numerical Integration
- Next by thread: Re: Numerical Integration
- Index(es):
Relevant Pages
|
Loading
