An easy looking integral, not so easy...
Easy looking integral, not so easy…
I saw this post from @infseriesbot and wanted to proof at least the first one. After a hardwork to compute this easy looking integral, I found out that the result stated by @infseriesbot is incorrect! Exctly, this time @infseriesbot is wrong. Checking numerically confirms the computation here.
Proof
let
Now use the fact that
for
The second integral is a little trickier
Observe the following
dividing both sides by and integrating from to
Now we have to evaluate the three integrals on the RHS and we are done. I´ll state the value of each of the integrals and show the proof in the end.
The next integral is the same as so no need to proof it again.
Putting all together we get
Now summing the results of and we get the final result!
Checking numerically we have and , which agrees with WolframAlpha!
A Corollary
We just computed the integral
On the other hand we can show that this integral equals
and conclude that
Proof:
Expanding in Taylor series we get
Integrating by parts we get:
and therefore
Proof that
Let
The last integral we have already computed, it´s ! and it´s equal to , therefore
and
Now let´s proof the harder one!
Lets start with the substitution
Let in the last integral.
(1)
We now evaluate each of these three integrals separately
using the formula that
We get
To compute I will first evaluate the indefinite version and then plug the limits.
Now integrate by parts with and
To evaluate we again consider Integration by parts.
Let
And . To find we used the same process used to calculate
Which is precisely , the integral we are looking to evaluate!
Plugging back all the results in (1) we get
Apendix
We used
Proof:
Splitting the first sum between even and odd terms
Ricardo Albahari
Comments
Post a Comment