Welcome to NRICH.

 
ò0 p/2 x cos2n(x) dx


By Brad Rodgers on Wednesday, August 28, 2002 - 09:24 pm:

Is there a closed form evaluation for

ò0 p/2 x cos2n(x) dx,

for integer n in general? I can easily evaluate it for individual n, but I don't see the pattern...

Brad


By Yatir Halevi on Wednesday, August 28, 2002 - 11:01 pm:

I don't know if it helps,
but ò0 p/2xcos2n(x)dx+ò0 p/2xsin2n(x)dx seems to have a connection to:
ò0 p/2cos2n(x)dx which equals to:
ò0 p/2sin2n(x)dx

altough I can't pin-point it yet,

Yatir


By Brad Rodgers on Thursday, August 29, 2002 - 01:58 am:

It appears that (from your work, Yatir)

ò0 p/2x(cos2nx + sin2nx) dx = ((2n)!/n!2)4-n-1 p2,

and I'll have a go at proving this. There's obviously a link between this and the original integral, but in the latter, a rational term appears that I just can't seem to determine.

(Of course, if anyone proves this, and makes a decent conjecture about the rational term, it'd be fairly easy to prove by induction; the problem is making the decent conjecture)

Brad


By Ian Short on Thursday, August 29, 2002 - 10:39 am:

Sorry to butt in- Possibly by substituting y=p/2-x into ò0 p/2xsin2nxdx it becomes clear how to evaluate the integral in the previous message.

Ian


By Michael Doré on Thursday, August 29, 2002 - 04:37 pm:

Also, for the original integral In = ò0 p/2 x cos2nx dx you can derive a reduction formula.

Write cos2nx = cos2n-2x - cos2n-2x sin2x. So:

In = In-1 - ò0 p/2x cos2n-2x sin2x dx

Integrate the second term by parts taking u = x sin x and dv/dx = cos2n-2x sin x, i.e. v = -cos2n-1x/(2n-1). This gives

In = In-1 - 1/(2n*(2n-1)) - In/(2n-1)

i.e.

2n In = (2n-1) In-1 - 1/(2n)