n = 1 ∑ ∞ P n ( − 1 ) n + 1 p n + 1
Let p n be the n th prime number. Let P n be the product of the first n primes. Which choice describes the summation 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.
Wow, it's like ours brains were synchronized. :) Do you know if there is an exact value for either the alternating or absolute-valued series?
Log in to reply
I was typing the solution as well; now that I returned back here, I found three essentially identical solutions within the span of 12 minutes. Apparently sufficiently difficult problems usually have only one easy-to-find solution.
Log in to reply
Yes, that's probably true. In this case, I think Bertrand's postulate was unavoidable.
Log in to reply
@Brian Charlesworth – I wonder if you can use similar logic to find if summations like n = 1 ∑ ∞ P n p 2 n + 1 , n = 1 ∑ ∞ P n p n 2 , or n = 1 ∑ ∞ P n p n n , converge. I suspect you can for subsets of the primes where the function of n is linear (I'm not in a position to write anything on paper at the moment) but I'm not so sure about the other cases.
Log in to reply
@Trevor B. – In the second and third of these series did you mean p n 2 and p n n ? If so, then I would hazard a guess that the first two of these series converge but not the last one. We could also look for the greatest real number k such that ∑ n = 1 ∞ P n p n k converges. Definitely worth investigating ....
Log in to reply
@Brian Charlesworth – No, p n 2 means the n 2 -th prime (as opposed to p n 2 which is the square of the n -th prime). Likewise for p n n .
Log in to reply
@Ivan Koswara – Ah, o.k., thanks for clarifying that; I hadn't thought of those terms before. Just having a quick look; do you know what lim n → ∞ p n 2 p n 2 equals? I don't think it's 0 .
Log in to reply
@Brian Charlesworth – I'm pretty sure it's zero. For large n , p n ≈ n ln n , so just substitute and simplify to get lim n → ∞ ln n 2 . Note that it's only approximation, though, so it's not rigorous.
Log in to reply
@Ivan Koswara – O.k., thanks, that seems reasonable. The rate of convergence would be slow, but it would nevertheless get there in the limit.
Let f ( n ) = P n ( − 1 ) n + 1 p n + 1 . Now by Bertrand's postulate we know that for n ≥ 1 we have that p n + 1 < 2 p n . Thus
∣ f ( n ) ∣ = P n p n + 1 < P n 2 p n = P n − 1 2 for n > 1 .
Also, since p n ≥ 2 for all n we know that P n − 1 ≥ 2 n − 1 for all n > 1 . Thus
∣ f ( n ) ∣ < P n − 1 2 ≤ 2 n 4 for n ≥ 2 , and so
n = 1 ∑ ∞ ∣ f ( n ) ∣ = ∣ f ( 1 ) ∣ + n = 2 ∑ ∞ ∣ f ( n ) ∣ < 2 3 + 4 ∗ n = 2 ∑ ∞ ( 2 1 ) n = 2 3 + 4 ∗ 1 − 2 1 4 1 = 2 7 .
As we have now established an upper bound for n = 1 ∑ ∞ ∣ f ( n ) ∣ we can conclude that the given series is absolutely convergent.
(Note that we can establish a tighter upper bound by isolating more of the initial terms. Using a similar argument, if we isolate ∣ f ( 1 ) ∣ and ∣ f ( 2 ) ∣ we can establish an upper bound of 3 8 . However, for the purposes of this question the initial bound of 2 7 will suffice.)
For this problem, we will make use of Bertand's Postulate .
We will also use the fact that where c is a nonzero constant, n = 1 ∑ ∞ a n and c n = 1 ∑ ∞ a n have the same convergence properties.
Bertrand's postulate states that between a number n and the number 2 n , there exists at least one prime. A direct consequence of this is that between a prime and its double (noninclusive), there is at least one other prime. Thus, p n < p n + 1 < 2 p n .
For now, we will ignore the ( − 1 ) n term and determine if n = 1 ∑ ∞ P n p n + 1 is divergent. (Obviously, if ∑ a n is absolutely convergent, then the series formed when alternate terms are negated is also convergent.)
Now we can determine the convergence of n = 1 ∑ ∞ P n p n + 1 .
Bertrand's postulate says that p n + 1 < 2 p n , so we can state the following. n = 1 ∑ ∞ P n p n + 1 < 2 n = 1 ∑ ∞ P n p n = 2 n = 1 ∑ ∞ P n − 1 1 We can now say that if 2 n = 1 ∑ ∞ P n − 1 1 is convergent, then n = 1 ∑ ∞ P n p n + 1 is convergent as well.
Multiplying a series by a constant does not affect its convergence. We can modify our previous statement to say that if n = 1 ∑ ∞ P n − 1 1 is convergent, then n = 1 ∑ ∞ P n p n + 1 is convergent as well.
Now we will examine the convergence of n = 1 ∑ ∞ P n − 1 1 .
All prime numbers are greater than or equal to 2 , so we know that P k ≥ 2 k for all positive integer k . Thus, P k 1 ≤ 2 k 1 for all positive integer k . Thus, we can say the following.
n = 1 ∑ ∞ P n 1 ≤ n = 1 ∑ ∞ 2 n 1 = 1 By the very large chain of convergies that we have established in this proof, we have that n = 1 ∑ ∞ P n ( − 1 ) n + 1 p n + 1 < 1 , meaning that the series is absolutely convergent
Nice question, Trevor. It's funny how all three of us posted variations of the same solution almost simultaneously. :)
Problem Loading...
Note Loading...
Set Loading...
By Bertrand's postulate , p n + 1 < 2 p n for all natural n , thus P n p n + 1 < P n 2 p n = P n − 1 2 .
Next, P n + 1 = p n + 1 ⋅ P n > 2 P n . Together with P 1 = 2 , we can induct to show that P n ≥ 2 n for all natural n .
Putting them all together,
n = 1 ∑ ∞ ∣ ∣ ∣ ∣ P n ( − 1 ) n + 1 p n + 1 ∣ ∣ ∣ ∣ = n = 1 ∑ ∞ P n p n + 1 = P 1 p 2 + n = 2 ∑ ∞ P n p n + 1 < 2 3 + n = 2 ∑ ∞ P n − 1 2 ≤ 2 3 + n = 2 ∑ ∞ 2 n − 1 2 = 2 3 + n = 2 ∑ ∞ 2 2 − n = 2 3 + 2 = 2 7
Thus ∑ n = 1 ∞ ∣ ∣ ∣ P n ( − 1 ) n + 1 p n + 1 ∣ ∣ ∣ is bounded above. Additionally, it's clearly monotonically increasing, and thus it converges. This proves that ∑ n = 1 ∞ P n ( − 1 ) n + 1 p n + 1 converges absolutely .