k = 1 ∑ ∞ k ϕ ( k ) ln ( 1 − x k )
Let ϕ ( k ) denotes Euler's Totient Function .
For 0 < x < 1 , find the closed form for the series above.
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! Well done! Part (2) and (3) will follow in the next days.
Log in to reply
Ah, I can't wait. Summations involving arithmetical functions are always a fun to solve. Gauss and Euler(Dirichlet too) have made them so much interesting, I thank those great souls.
Do you have the proof for Gauss' argument?
And can you type the non-English (Dirichlet convolution) solution as well?
Log in to reply
That is a classical formula, which can be proved in the following simple way: consider the rational numbers with denominator n and numerator ranging from 1 to n.When reducing them, we get exactly φ ( d ) numbers with denominator d|n.This is true, because exactly φ ( d ) of the numbers d 1 , d 2 , . . . , d d are reduced.Since every reduced fraction is uniquely represented, this shows that d ∣ n ∑ φ ( d ) = n
Log in to reply
Ohhh! That simple? Thanks!
Log in to reply
@Pi Han Goh – No problem :D.There are of course proofs that offer more insight (one can prove it also by using the identity Φ n ( x ) = d ∣ n ∏ Φ d ( x ) )
Log in to reply
@Bogdan Simeonov – I haven't mastered cyclotomic polynomials so I'm a little fuzzy in it. Thanks again!
I do miss your questions and answers though. Hope you can post some sometime soon.
Log in to reply
@Pi Han Goh – Well, maybe I will get back to it sometime.Recently I got pretty interested in abstract algebra so maybe brilliant could expand to more advanced topics.For now I browse through math.stackexchange sometimes, because there's a myriad of great problems (and really smart people) there .
Let f ( x ) = k = 1 ∑ ∞ k ϕ ( k ) ln ( 1 − x k ) This implies f ′ ( x ) = k = 1 ∑ ∞ k ϕ ( k ) 1 − x k − k x k − 1 = − x 1 k = 1 ∑ ∞ 1 − x k ϕ ( k ) x k Using Lambert series which states for ∣ x ∣ < 1 k = 1 ∑ ∞ 1 − x k ϕ ( k ) x k = ( 1 − x ) 2 x Thus f ′ ( x ) = − ( 1 − x ) 2 1 ⟹ f ( x ) = − 1 − x 1 + C We evaluate C by plugging in x = 0 to get C = 1 and finally f ( x ) = − 1 − x 1 + 1 = 1 − x − x
Great observation!
Log in to reply
I did find a closed form for 0 < x < 1 but the series is still valid for x = 0 .
Why do you see subbing x = 0 as a problem?
Problem Loading...
Note Loading...
Set Loading...
k = 1 ∑ ∞ k ϕ ( k ) ln ( 1 − x k ) = − k = 1 ∑ ∞ k ϕ ( k ) ( − ln ( 1 − x k ) )
− k = 1 ∑ ∞ k ϕ ( k ) m = 1 ∑ ∞ m x k m
Now, you can just use Dirichlet series by utilizing Dirichlet convolution. But we'll do this in a beginner's way. We'll write our series in the following way:
− n = 1 ∑ ∞ a ∣ n ∑ a m ϕ ( a ) x n
where a m = n so it simply becomes
− n = 1 ∑ ∞ a ∣ n ∑ ϕ ( a ) n x n
Now, using Gauss' argument that a ∣ n ∑ ϕ ( a ) = n [will not prove here]
= − n = 1 ∑ ∞ x n = 1 − x − x