k = 1 ∑ ∞ n = 0 ∑ ∞ k ( 1 + 2 n k ) ( − 1 ) k − 1
Find the value (to 2 decimal places) of the closed form of the expression 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.
Oh wow, that's a really innovative approach. What made you think of A and B , and to find the identity between them?
Log in to reply
Writing an alternating sum as "the sum of the lot, minus twice the sum over the even-indexed terms" is a fairly standard ploy. B is the sum over all even k .
Log in to reply
Ah, it didn't occur to me immediately that B was the "negative portion". Could you add this line to the top of the solution to provide the motication of what you did? Thanks!
Note that the sum (say S ) converges absolutely as k ( 1 + k 2 n ) 1 < k 2 2 n 1 . Hence, we can interchange the order of summations. Interchanging and writing 1 + k 2 n 1 as ∫ 0 1 x k 2 n d x , we get the following :
S = n = 0 ∑ ∞ k = 1 ∑ ∞ k ( − 1 ) k − 1 ∫ 0 1 x k 2 n d x = n = 0 ∑ ∞ ∫ 0 1 ln ( 1 + x 2 n ) d x ( by uniform convergence of power series ) = ∫ 0 1 ln ( 1 − x 1 ) d x ( by Lebesgue’s monotone convergence theorem ) = − ∫ 0 1 ln x d x = 1
Note: The following equation is also used to get the third equality : n = 0 ∏ N − 1 ( 1 + x 2 n ) = 1 − x 1 − x 2 N
You say that the first identity is due to the uniform convergence of power series. Yes and no - you need to be more explicit. While it is true that power series converge uniformly on compact subsets of their (open) interval of convergence (i.e for ∣ x ∣ ≤ S for any S < R , the radius of convergence), we cannot use this result here, since we are integrating on [ 0 , 1 ] , where R = 1 .
We have to use the alternating nature of this series, plus the fact that the terms in the sequence have decreasing modulus, so that ∣ ∣ ∣ ∣ ∣ ln ( 1 + x 2 n ) − k = 1 ∑ K k ( − 1 ) k − 1 x k 2 n ∣ ∣ ∣ ∣ ∣ ≤ k + 1 1 x ( k + 1 ) 2 n ≤ k + 1 1 0 ≤ x ≤ 1 to show that the series converges uniformly on [ 0 , 1 ] .
Log in to reply
(Sorry for the really late reply)
We could interchange the sum and integral over [ 0 , r ] for r < 1 and then use the continuity of F ( x ) = ∫ 0 x f ( t ) d t to conclude that the same can be done over [ 0 , 1 ] right?
P.S.: Mostly this is wrong, but I want to confirm.
Log in to reply
That should work, but it's a bit fiddly...
Log in to reply
@Mark Hennings – Hmm... Okay.
Thanks for the reply.
Thanks for justifying all of those steps. At the Putnam, you have to explain and justify your work carefully.
Let S = k = 1 ∑ ∞ n = 0 ∑ ∞ k ( k ⋅ 2 n + 1 ) ( − 1 ) k − 1
Since S is absolutely convergent, we can swap summations to get,
S = n = 0 ∑ ∞ k = 1 ∑ ∞ k ( k ⋅ 2 n + 1 ) ( − 1 ) k − 1
= n = 0 ∑ ∞ k = 1 ∑ ∞ [ k ( − 1 ) k − 1 − k + 2 n 1 ( − 1 ) k − 1 ]
Splitting into even and odd terms, we have,
= n = 0 ∑ ∞ k = 1 ∑ ∞ [ k ( − 1 ) k − 1 − ( k + 2 n 1 1 − k + 2 n + 1 1 1 ) ]
Converting to Digamma Function , we have,
= n = 0 ∑ ∞ [ lo g 2 + ψ ( 2 n 1 ) − ψ ( 2 n + 1 1 ) − 2 n ]
= N → ∞ lim n = 0 ∑ N [ lo g 2 + ψ ( 2 n 1 ) − ψ ( 2 n + 1 1 ) − 2 n ] ( ∗ )
Clearly, ( ∗ ) telescopes. Evaluating it, we have,
S = 1
After splitting, shouldn't the exponent of 2 in the end be n + 1 rather than n − 1 ?
Let S0 and S1 be the
sums ∑k
1
k ∑
∞
n=0
1
k2
n+1 with k running over all odd and
all even positive integers, respectively, so that
S = S1 −S0 = (S0 +S1)−2S0.
In S1, we may write k = 2
to obtain
S1 =
∞
∑
=1
1
2
∞
∑
n=0
1
2
1 2 (S0 +S1) + ∞ ∑ `=1 1
(
+1)
1 2 (S0 +S1) + 1 2 because the last sum telescopes; this immediately yields S = 1.
Nice solution.I solved it the harder way.
Problem Loading...
Note Loading...
Set Loading...
Let A = k = 1 ∑ ∞ n = 0 ∑ ∞ k ( 2 n k + 1 ) 1 B = k = 1 ∑ ∞ n = 0 ∑ ∞ 2 k ( 2 n + 1 k + 1 ) 1 Note that while A is the sum of k ( 2 n k + 1 ) 1 over all n ≥ 0 and k ≥ 1 , the term B is the sum of k ( 2 n k + 1 ) 1 over all n ≥ 0 and even k ≥ 1 . Now B = 2 1 k = 1 ∑ ∞ n = 1 ∑ ∞ k ( 2 n k + 1 ) 1 = 2 1 [ k = 1 ∑ ∞ n = 0 ∑ ∞ k ( 2 n k + 1 ) 1 − k ( k + 1 ) 1 ] = 2 1 ( A − 1 ) But the sum we are interested in is k = 1 ∑ ∞ n = 0 ∑ ∞ k ( 2 n k + 1 ) ( − 1 ) k − 1 = A − 2 B = 1