∫ 0 ∞ x e − 2 x sin x d x = a π − arctan b
The integral above holds true for the integers a and b . Find a + b .
This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try
refreshing the page, (b) enabling javascript if it is disabled on your browser and,
finally, (c)
loading the
non-javascript version of this page
. We're sorry about the hassle.
Thanks for your solution Sir. But first line have a typo, denominator should be x not 2 .
Nice solution. As I mentioned in Kelvin's comment thread, though, you have a note of " I ( ∞ ) = 0 ", which is true but nontrivial to show.
Relevant Wiki: Differentiation Under the Integral Sign
Let f ( t ) = ∫ 0 ∞ x e − t x sin x d x , by Differentiation under the integral sign: f ′ ( t ) = ∫ 0 ∞ ∂ t ∂ x e − t x sin x d x = ∫ 0 ∞ − e − t x sin x d x
By using integration by part, we have the table below:
sin x cos x − sin x − e − t x t 1 e − t x − t 2 1 e − t x
So we ended up with
f ′ ( t ) ( 1 + t 2 ) f ′ ( t ) f ′ ( t ) ∴ f ( t ) = t e − t x sin x + t 2 e − t x cos x − t 2 1 f ′ ( t ) = e − t x ( t sin x + cos x ) = 1 + t 2 t sin x + cos x e − t x ∣ ∣ ∣ ∣ ∣ 0 ∞ = − 1 + t 2 1 = − arctan t + C
To calculate C , we may consider f ( ∞ ) , so
f ( ∞ ) = t → ∞ lim ∫ 0 ∞ x e − t x sin x d x = 0 which help us to determine that C = 2 π Compute t = 2 , we now have ∫ 0 ∞ x e − 2 x sin x d x = f ( 2 ) = 2 π − arctan 2
@Kelvin Hong How would you explain f ( ∞ ) = 0 ?
Log in to reply
From the original integral, t goes infinity then exponential goes to zero.
Log in to reply
Oh okay so you mean that ∫ 0 ∞ x s i n x d x = 0 but why? Can it be assumed like that?
@donglin loo No, e − ∞ actually goes to zero, so it is same as calculate ∫ 0 ∞ 0 d x . Also, you can see in page 4 in this pdf file http://www.math.uconn.edu/~kconrad/blurbs/analysis/diffunderint.pdf That ∫ 0 ∞ x sin x d x = 2 π .
Log in to reply
@Kelvin Hong – I see. Thanks for that clarification.
@Kelvin Hong – Note that we cannot in general interchange limits with integrals. Your justification essentially is that the integrand goes to zero, but to show the integral goes to zero is another thing entirely.
To be clear, your conclusion is correct, but it's a nontrivial exercise to show that the limit of the integrals is also zero.
Log in to reply
@Brian Moehring – I wonder why is it non-trivial? It is like finding the area between the line y = 0 and x axis from zero to infinity, it is zero right?
Log in to reply
@Kelvin Hong – Oh, I'm not saying that it's nontrivial to show ∫ 0 ∞ 0 d x = 0 -- that is trivial. The problem is that you don't have that. You have a limit of integrals whose integrands go to zero. Like differentiating under the integral, it requires something more the just the limit existing to do it.
In particular, I am saying that this type of step is nontrivial: t → ∞ lim ∫ 0 ∞ x e − t x sin x d x = ∫ 0 ∞ ( t → ∞ lim x e − t x sin x ) d x
The easiest way to justify it (as far as I can tell... I may be missing something) is to note ∣ ∣ ∣ ∣ x e − t x sin x ∣ ∣ ∣ ∣ ≤ e − x for all t ≥ 1 and that ∫ 0 ∞ e − x d x = 1 < ∞ . This is the set-up for Lebesgue's Dominated Convergence Theorem (which is probably the most celebrated theorem allowing us to interchange limits and integrals)
Log in to reply
@Brian Moehring – Wow, I didn't think of that aspect, thanks for the information !
we can use f(0) = pi/2 as it is the common dirichlet integral .
Log in to reply
We can use this to find C , but note that the given solution only works for t > 0 , so in order to use that fact, we would have to show the integral defining f is continuous up to t = 0 .
That is, you would have to justify a limit of an integral, and as far as I can tell, the required limit would be harder than the one used in Kelvin's solution.
Log in to reply
what if we just apply lim t-> 0+ on f(t). Then the solution would work . Only using the right hand limit would give the same result
Log in to reply
@Arghyadeep Chatterjee – That's what I was saying. To be more clear, you would need to justify t → 0 + lim ∫ 0 ∞ x e − t x sin x d x = ∫ 0 ∞ x sin x d x or equivalently, since we already know the right integral is finite, t → 0 + lim ∫ 0 ∞ x ( 1 − e − t x ) sin x d x = 0
While I was able to find an easy justification for the limit required by Kelvin's solution, I can't think of one for this limit (though we can confirm it's true indirectly)
∫ 0 ∞ x e − 2 x S i n x d x = a π − a r c t a n ( b ) This integration can be evaluated by exploiting Feynman's trick. Considering the parameterisation ⟹ I ( a ) = ∫ 0 ∞ x S i n x e − a x d x a ≥ 0 ⟹ I ( a ) = − ∫ 0 ∞ e − a x S i n x d x
Differentiating w.r.t x ⟹ I ′ ( a ) = − 1 + a 2 1 Note that, since we are interested in I ( 2 ) and we got I ( 0 ) = 2 π
We obtain that, − ∫ 0 2 1 + a 2 1 d a = − [ I ( 2 ) − I ( 0 ) ] = − [ a r c t a n ( 2 ) 2 π ] ⟹ 2 π − a r c t a n ( 2 ) = a r c t a n ( 2 1 )
The last equality follows from the relationship between the inverse tangent and inverse co-tangent function. The value of a + b = 4
Problem Loading...
Note Loading...
Set Loading...
Relevant wiki: Differentiation Under the Integral Sign
Similar solution with @Kelvin Hong's using differentiation under the integral sign.
I ( a ) ∂ a ∂ I ( a ) ⟹ I ( a ) I ( ∞ ) ⟹ I ( 2 ) = ∫ 0 ∞ x e − a x sin x d x = − ∫ 0 ∞ e − a x sin x d x = e − a x cos x ∣ ∣ ∣ ∣ 0 ∞ + a ∫ 0 ∞ e − a x cos x d x = 0 − 1 + a e − a x sin x ∣ ∣ ∣ ∣ 0 ∞ + a 2 ∫ 0 ∞ e − a x sin x d x = − 1 + 0 − 0 − a 2 ⋅ ∂ a ∂ I ( a ) = − a 2 + 1 1 = − ∫ a 2 + 1 1 d a = − arctan a + C = − arctan ∞ + C = 0 = − 2 π + C = 0 = 2 π − arctan 2 Differentiate w.r.t. a . By integration by parts Note that ∂ a ∂ I ( a ) = − ∫ 0 ∞ e − a x sin x d x where C is the constant of integration. Note that I ( ∞ ) = 0 ⟹ C = 2 π Note that I ( 2 ) is the required integral.
Therefore, a + b = 2 + 2 = 4 .