∫ 0 1 1 − x 2 ln ( x 4 − 2 x 2 + 5 ) − ln ( 5 x 4 − 2 x 2 + 1 ) d x
Given that the integral above is equal to π α arctan δ β − γ , where α , β , γ , δ are integers with β square-free.
Calculate α + β + γ + δ .
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.
This is the most number of substitutions I've ever seen!
Question: What motivates you to substitute y = 1 − x 2 1 + x 2 at the very beginning? That substitution appears out of nowhere to me and I don't understand the reasoning behind this substitution.
want to know the motive
I too don't get it. how did you arrive at the first substitution?
It is a natural substitution as it allows the IBP and DUI to become more effectively noticeable
Picking up quite early on in Ronak's proof, we have I = 2 ∫ 1 ∞ y 4 + 4 y 2 − 1 ln ( y + y 2 − 1 ) , d y . Substituting y = cosh u , and later v = e u gives the integral I = = = = 2 ∫ 0 ∞ cosh 4 u + 4 cosh 2 u − 2 u sinh u d u = ∫ R cosh 4 u + 4 cosh 2 u − 2 u sinh u d u 2 ∫ R ( e u + e − u ) 4 + 6 4 ( ( e u + e − u ) 2 − 8 ) ( e u − e − u ) u d u 2 ∫ R ( e 2 u + 1 ) 4 + 6 4 e 4 u ( ( e 2 u + 1 ) 2 − 8 e 2 u ) ( e 2 u − 1 ) u e u d u 2 ∫ 0 ∞ ( v 2 + 1 ) 4 + 6 4 v 4 ( ( v 2 + 1 ) 2 − 8 v 2 ) ( v 2 − 1 ) ln v d v = 2 ∫ 0 ∞ F ( v ) ln v d v . Now F is a rational function of v whose denominator has degree 2 more than its numerator. If we take the principal branch of the logarithm, cutting the plane along the negative real axis, and consider the contour which runs
then standard contour integration (letting R → ∞ and ϵ → 0 ) tells us that I + π i ∫ 0 ∞ F ( v ) d v = 2 π i q ∈ R ( + ) ∑ R e s z = q F ( z ) lo g z , where R ( + ) is the set of poles of F ( z ) with positive imaginary part. The order 8 polynomial which is the denominator of F can be factorized quite easily, and we obtain the fact that R ( + ) = { a , b , c , d } , where a c = = − 1 + i − − 1 − 2 i − 1 − i + − 1 + 2 i b d = = 1 − i − − 1 − 2 i 1 + i + − 1 + 2 i (the square roots are chosen with positive real part), and we note that a = − d ⋆ and c = − b ⋆ . From here on in it is just number-crunching. The sum of the residues is q ∈ R ( + ) ∑ R e s z = q F ( z ) lo g z = = = 8 1 [ ( lo g d − lo g a ) + ( lo g b − lo g c ) ] 8 1 i [ δ − ( π − δ ) + β − ( π − β ) ] 4 1 i ( β + δ − π ) (since ∣ a ∣ = ∣ d ∣ and ∣ b ∣ = ∣ c ∣ ), from which we deduce that I = 2 1 π ( π − β − δ ) where 0 ≤ β , δ ≤ 2 1 π are the arguments of b , d respectively. After a fair amount of trigonometric calculation, we can show that tan 2 1 ( β + δ ) = 2 5 + 1 , and hence I = π ( 2 1 π − 2 1 ( β + δ ) ) = π tan − 1 2 5 − 1 , as required.
This is still not a perfect solution, just an alternative one. I feel pretty certain that a more subtle piece of contour integration will get the answer more efficiently...
Problem Loading...
Note Loading...
Set Loading...
∫ 0 1 1 − x 2 ln ( x 4 − 2 x 2 + 5 ) − ln ( 5 x 4 − 2 x 2 + 1 ) d x = π arctan 2 5 − 1
The method I am going to use is not the best method to do it hence any help in simplification is appreciated.
Proof :
I = ∫ 0 1 1 − x 2 lo g ( x 4 − 2 x 2 + 5 ) − lo g ( 5 x 4 − 2 x 2 + 1 ) d x
Start with the substitution y = 1 − x 2 1 + x 2 , to get :
I = 2 1 ∫ 1 ∞ y 2 − 1 ln ( y 2 − 2 y + 2 y 2 + 2 y + 2 ) d x
Integrating by parts : ∫ u v ′ = u v − ∫ u ′ v , u = ln ( y 2 − 2 y + 2 y 2 + 2 y + 2 ) , v = y 2 − 1 1 we get :
I = ln ( y 2 − 2 y + 2 y 2 + 2 y + 2 ) ln ( y + y 2 − 1 ) ∣ 1 ∞ + 2 ∫ 1 ∞ y 4 + 4 y 2 − 2 ln ( y + y 2 − 1 ) d x
⇒ I = 2 ∫ 1 ∞ y 4 + 4 y 2 − 2 ln ( y + y 2 − 1 ) d x
Now we introduce a parameter here :
f ( a ) = a ∫ 1 ∞ y 4 + a 4 y 2 − a 2 ln ( y + y 2 − 1 ) d x
Put y = a x , to get :
f ( a ) = ∫ 1 / a ∞ x 4 + 1 x 2 − 1 ln ( a x + ( a x ) 2 − 1 ) d x
Differentiate with respect to a we have and using newton leibnitz rule :
f ′ ( a ) = ( a 1 ) 2 ( 1 / a ) 4 + 1 ( 1 / a ) 2 − 1 ln ⎝ ⎛ a ( a 1 ) + ( a ( a 1 ) ) 2 − 1 ⎠ ⎞ + ∫ 1 / a ∞ x 4 + 1 x 2 − 1 ( a x ) 2 − 1 x d x
⇒ f ′ ( a ) = ∫ 1 / a ∞ x 4 + 1 x 2 − 1 ( a x ) 2 − 1 x d x
Put ( a x ) 2 − 1 = y 2 to get :
f ′ ( a ) = ∫ 0 ∞ ( y 2 + 1 ) 2 + a 4 ( y 2 + 1 − a 2 ) d y
With the substitution y ↦ y a 4 + 1 we get :
f ′ ( a ) = a 4 + 1 1 ∫ 0 ∞ y 4 + 2 y 2 + ( a 2 + 1 ) ( 1 − a 2 ) y 2 + ( a 4 + 1 ) d y
Adding these two we have :
f ′ ( a ) = ( 2 a 4 + 1 ( 1 − a 2 ) + a 4 + 1 ) ∫ 0 ∞ y 4 + 2 y 2 + ( a 2 + 1 ) y 2 + a 4 + 1 d y
f ′ ( a ) = ( 2 a 4 + 1 ( 1 − a 2 ) + a 4 + 1 ) ∫ 0 ∞ ( y − y a 4 + 1 ) 2 + 2 ( 1 + a 4 + 1 ) 1 + y 2 a 4 + 1 d y
Use the substitution ( y − y a 4 + 1 = x ) to get :
f ′ ( a ) = ( 2 a 4 + 1 ( 1 − a 2 ) + a 4 + 1 ) ∫ − ∞ ∞ x 2 + 2 ( 1 + a 4 + 1 ) d x
⇒ f ′ ( a ) = ( 2 a 4 + 1 ( 1 − a 2 ) + a 4 + 1 ) 2 ( 1 + a 4 + 1 ) π
Simplifying it we have :
f ′ ( a ) = 2 π ( a 4 + 1 1 ) 2 ( 1 + a 4 + 1 ) 2 ( a 4 + 1 ) − 2 a 2 + 2 ( 1 − a 2 ) a 4 + 1
⇒ f ′ ( a ) = 2 π a 4 + 1 a 4 + 1 − a 2
Now integrating both sides we have :
∫ 0 a f ′ ( x ) d x = 2 π ∫ 0 a x 4 + 1 x 4 + 1 − x 2 d x
We will evaluate G = ∫ 0 a x 4 + 1 x 4 + 1 − x 2 d x
Put x 4 + 1 = y + x 2 , to get :
G = 2 1 ∫ a 4 + 1 − a 2 1 1 − y 2 d y
⇒ G = 2 2 arccos ( a 4 + 1 − a 2 )
Hence G = 2 arctan ⎝ ⎛ a 2 a 4 + 1 − 1 ⎠ ⎞
Finally we have :
f ( a ) = f ( 0 ) + 2 π arctan ⎝ ⎛ a 2 a 4 + 1 − 1 ⎠ ⎞
It is obvoius that f ( 0 ) = 0
Finally we have our result :
f ( a ) = 2 π arctan ⎝ ⎛ a 2 a 4 + 1 − 1 ⎠ ⎞
For solving our problem we know that I = 2 f ( 2 )
⇒ I = π arctan ⎝ ⎜ ⎛ 2 5 − 1 ⎠ ⎟ ⎞