∫
2
3
ln
(
Γ
(
x
)
)
d
x
If the result of the integral above can be expressed in the form
c
a
ln
c
+
c
b
ln
π
−
c
where
a
,
b
and
c
are positive coprime integers. Find the value of
a
+
b
+
c
.
Note that Γ ( x ) is the Gamma Function.
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.
Great! Thanks for proving the complicated result.
Hats off to u! Seriously that may have given u a headache.
The integral is a Raabe's integral named after Joseph Ludwig Raabe and it is given by:
∫ a a + 1 ln ( Γ ( x ) ) d x ⇒ ∫ 2 3 ln ( Γ ( x ) ) d x ⇒ a + b + c = 2 1 ln 2 π + a ln a − a = 2 1 ln 2 π + 2 ln 2 − 2 = 2 1 ln 2 + 2 1 ln π + 2 ln 2 − 2 = 2 5 ln 2 + 2 1 ln π − 2 = 5 + 1 + 2 = 8
Yes, that's the result to use. It takes a bit of work to demonstrate.
Sir, can you prove the result of the integral you mentioned in the beginning? Raabe's integral?
Log in to reply
Well, I can give you some insights of how to prove Raabe's theorem and may give the full proof too if I get time. So, for now, only some hints.
Use the Weierstrass product for gamma function
Γ ( z ) = z e − γ z n = 1 ∏ ∞ ( 1 + n z ) e − z / n
Now take the logarithm both sides, then do the integration, of course, from a to a + 1 and you are done. You may also need some knowledge of ζ ′ ( s ) or even the derivative of Hurwitz zeta function(which I guess might come).
Log in to reply
Thanks, Kartik.
Log in to reply
@Chew-Seong Cheong – Sir, you could also consider I ( s ) = ∫ a a + 1 ln ( Γ ( s z ) ) d z and the use Differentiation Under the Integral Sign.
Log in to reply
@User 123 – I don't see how you will go ahead. I too tried it but it doesn't make any big difference. Some more digammas won't help, I guess.
Log in to reply
@Kartik Sharma – I don't know if it works out or not. I just evaluated I ′ ( s ) and then left it. By the way, how did you solve Summations can be Easy ? I would be really grateful if you could please show your method at your convenience.
Log in to reply
@User 123 – Well,digamma to the rescue. I had learnt in Abramowitz and Stegun how to directly transform such sums into digamma(well, although there is nothing to learn, it becomes so obvious once you have done some success in simplifying, but I must "honor" that book because it is a marvel(if not for this particular case, in general))
Log in to reply
@Kartik Sharma – Thanks! Would you know from where I can get hold of the a PDF of that book?
Log in to reply
@Kartik Sharma You could also do this: ∫ a a + 1 lo g Γ ( t ) d t = n → + ∞ lim k = 0 ∑ n − 1 lo g Γ ( a + n k ) = n → + ∞ lim n 1 lo g k = 0 ∏ n − 1 Γ ( a + n k ) and then apply the Multiplication Theorem . This one is courtesy of my Uncle:)
Log in to reply
@User 123 – Well, that is really good. I liked it. I had been too straight at the proof and was losing some details. I am a fan of your uncle!
Problem Loading...
Note Loading...
Set Loading...
I = ∫ a a + 1 ln Γ ( x ) d x
Apply integration by parts with u = ln Γ ( x ) and d v = d x , and use the definition of the Digamma Function :
I = x ln Γ ( x ) ∣ a a + 1 − ∫ a a + 1 x ψ ( x ) d x
Recall the Harmonic Number - Digamma Function Relation:
ψ ( x ) = H x − 1 − γ = H x − x 1 − γ ,
where γ is the Euler-Mascheroni Constant .
= > I = a ln a + ln a ! − ∫ a a + 1 x ( H x − x 1 − γ ) d x
= a ln a + ln a ! + 1 + 2 2 a + 1 γ − ∫ a a + 1 x H x d x
= a ln a + ln a ! + 1 + 2 2 a + 1 γ − ∫ a a + 1 x k = 1 ∑ ∞ k ( x + k ) x d x
= a ln a + ln a ! + 1 + 2 2 a + 1 γ − k = 1 ∑ ∞ ∫ a a + 1 k ( x + k ) x 2 d x
this follows from the definition of the Harmonic Number for all real values. Now let's work out the last term alone: this integral can be evaluated easily using basic integration techniques, to get at last:
k = 1 ∑ ∞ ∫ a a + 1 k ( x + k ) x 2 d x
= k = 1 ∑ ∞ ( 2 2 a + 1 k 1 − 1 + k ln ( a + k + 1 ) − k ln ( a + k ) )
= n → ∞ lim k = 1 ∑ n ( 2 2 a + 1 k 1 − 1 + ( k + 1 ) ln ( a + k + 1 ) − k ln ( a + k ) − ln ( a + k + 1 ) )
n → ∞ lim ( 2 2 a + 1 H n − n + ( n + 1 ) ln ( a + n + 1 ) − ln ( a + 1 ) − ln ( a + 1 ) ! ( a + n + 1 ) ! )
Rearrange the logarithms into one log, with n = ln e n :
= n → ∞ lim ( 2 2 a + 1 H n + ln ( ( a + n + 1 ) ! e n ( a + n + 1 ) n + 1 a ! ) )
Use Stirling's formula for the factorial in the log ( as we have n → ∞ ). After simplification, we get:
= n → ∞ lim ( 2 2 a + 1 H n + ln ( ( a + n + 1 ) a + 2 1 2 π e a + 1 a ! ) )
= n → ∞ lim ( 2 2 a + 1 H n − 2 2 a + 1 ln ( n + a + 1 ) ) + ln ( 2 π e a + 1 a ! )
Recall the definition of the Euler-Mascheroni Constant:
= 2 2 a + 1 γ + a + 1 + ln a ! − 2 1 ln ( 2 π )
Now substitute this back in our expression of I :
I = a ln a + ln a ! + 1 + 2 2 a + 1 γ − k = 1 ∑ ∞ ∫ a a + 1 k ( x + k ) x 2 d x
= a ln a + ln a ! + 1 + 2 2 a + 1 γ − ( 2 2 a + 1 γ + a + 1 + ln a ! − 2 1 ln ( 2 π ) )
I = 2 1 ln ( 2 π ) + a ln a − a
Let a = 2 , we get our desired answer to this problem.