Re: Directional derivative using finite differences
- From: Mario S. Mommer <m_mommer@xxxxxxxxx>
- Date: 31 Mar 2005 06:37:53 +0000
Hi,
"A Net User" <matlabber@xxxxxxxxx> writes:
> How can one find a directional derivative using finite differences?
You probably mean 'approximate'.
> For example considering a grid as below with a grid size of 1
>
> a b c
> d e f
> g h i
>
> the derivative along x direction at point 'e' can be found as (f-d)/2
> using central differences and second order derivative would be
> (f+d-2e).
>
> What about the derivative at point 'e' along a diagonal line passing
> through points 'g','e' and 'c'?
Write them in terms of the grid size h. By setting your grid size to
h=1, you have just hidden the fact that you can use the same formula
(with a different h) for that particular derivative :-)
You can also approximate any directional derivative by using the
formula
dF dF dF
-- = ( -- , -- ) v, v in R^2 a unit vector.
dv dx dy
and then compute dF/dx, etc. using finite differences.
.
- References:
- Directional derivative using finite differences
- From: A Net User
- Directional derivative using finite differences
- Prev by Date: A question on Newton's Method
- Next by Date: Re: A question on Newton's Method
- Previous by thread: Directional derivative using finite differences
- Next by thread: A question on Newton's Method
- Index(es):
Relevant Pages
|
