|
|
 | | From: | Friest Welheim | | Subject: | Integration Problem | | Date: | Sun, 23 Jan 2005 22:42:06 -0500 |
|
|
 | Does anyone know how to get an accurate integration from 0 to 0.2 of
y = exp(1.2*x)*besselj(0,2*sqrt(x))
I have tried using int but it failed to provide me with an explicit
results. Quad and quadl could not produce the necessary accuracy. Any
suggestions?
|
|
| | From: | Roger Stafford | | Subject: | Re: Integration Problem | | Date: | Mon, 24 Jan 2005 09:19:21 GMT |
|
|
| In article , "Friest Welheim"
wrote:
> Does anyone know how to get an accurate integration from 0 to 0.2 of
>
> y = exp(1.2*x)*besselj(0,2*sqrt(x))
>
> I have tried using int but it failed to provide me with an explicit
> results. Quad and quadl could not produce the necessary accuracy. Any
> suggestions?
-------
Very likely your trouble is due to the behavior of sqrt(x) as x
approaches zero where its derivative approaches infinity. Try a change of
variable u = sqrt(x) so that you are then integrating
y = exp(1.2*u^2)*besselj(0,2*u)*2*u
with respect to u from u = 0 to u = sqrt(.02). Then quad or quadl should
give you good accuracy even if int fails to find an explicit solution
--
(Remove "xyzzy" and ".invalid" to send me email.)
Roger Stafford
|
|
|
