That was my first thought...but try this formula on your result

(u/v)''=(u''v-v''u)/(v^2)

Ummm... seems to doesn''t work

the u/v thing is when differentiating.

the differential of cos nx is -nsin nx, so the integral of
sin nx is (-cos nx)/n

(Where n is a constant)

I don''t have time to go into the laws of integration

i have tried to integrate that function in mathcad and the result
had some fourier transorm shit in it and pi raised to some weird power, this is a very complex intergration

My TI-89 won''t even try to integrate that. Good luck.

mm... i thought this was simple. I''ll go over it again lter on paper - I cannot think properly without a ''scratchpad''

my notes in class look awful

ok, first off

the integral of sin(x^2) IS NOT (-cos(x^2))/(2x), you may be able to differentiate back to sin(x^2) but (-cos(x^2))/(2x) does not exist at 0 but the integral of sin(x^2) at 0 does. (check: expand sin(x^2) as a maclurin series and then integrate .. 0 does exist). Also if i remember correctly sin(x^2) does not have integral within the basic functions of calculus ... so you must expand it as a Taylor series then integrate ...

Second,

I would go with declspec suggestion and expand sin(x^2) as maclurin series then intergrate. Now it's up to you on how many terms you need, more terms = more accurate approximation of intergral of sin(x^2).

hope that helped.
jp.



==============================================
I feel like a kid in some kind of store...
==============================================
www.thejpsystem.com

[edited by - loserkid on May 28, 2002 4:11:03 PM]

Mine was analytic. Not that mine was the only analytic one.

[edited by - LilBudyWizer on May 24, 2002 2:34:42 PM]

Thanks guys, this has been informative. As it turns out my teacher is going to let me do the integral with a maclauren series expansion.. so that''ll make it nice and clean. And as it turns out for the required domain I only need like 4 terms of the series to be accurate enough.

Thanks!

Basically, if you find the integral of f(x), but when differentiating back to f(x) you cancel out an x term... that usually means you don''t have the proper integral in the first place. Cancelling out x terms is sometimes bad, especially since x/x is indeterminate when x = 0. Watch out!

As many people have said the integral ₀∫^a sin(x²) dx is an elementary function called the Fresnel sin function. These are evaluated numerically and look quite pretty when plotted.

For the specific case ₀∫^infinity sin(x²) dx you can actually evaluate it.

Make the change of variables u = x^2 and then use the formula sin(u) = 0.5*i(e^-iu - e^iu)

You then make two more changes of variables on the two integrands containing e^-iu and e^iu and you end up with this

0.5 * cos(PI/4) * ₀∫^infinityv^-0.5e^-v dv

The integral on the right is Gamma(0.5) = sqrt(PI) where Gamma is the generalised factorial function.

So the whole thing is 0.5 * sqrt(PI/2).

Don't know if this helps but it's kinda neat


[edited by - sQuid on May 28, 2002 11:33:47 PM]