Amazing prime numbers!

Can we divide the first 2019 prime numbers into two groups such that the sum of the two groups are equal?

No Yes It depends

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.

1 solution

Mark Hennings
Feb 13, 2020

If the n n th prime number is p n p_n , then p 36 + j = 68 1461 p j = 8303321 = 1 2 j = 1 2019 p j p_{36} + \sum_{j=68}^{1461}p_j \; = \; 8303321 \; = \; \frac12\sum_{j=1}^{2019}p_j so the split is possible.


Now for detail...


Proposition Any positive integer N 7 N \ge 7 can be written as the sum of distinct positive integers, none of which are greater than m a x ( 11 , N 7 ) \mathrm{max}(11,N-7) .

Simple checking shows that this works for all 7 N 27 7 \le N \le 27 . Now suppose that N 28 N \ge 28 , and that this result holds for all integers n n between 7 7 and N 1 N-1 . If we set M = N 6 2 M \; = \; \left\lfloor \frac{N-6}{2}\right\rfloor then M > 10 M > 10 and so Bertrand's (no longer a) Postulate tells us that there exists a prime number p p such that M + 1 p 2 M 1 M+1 \le p \le 2M-1 . But then p 2 M 1 2 ( N 6 2 ) 1 = N 7 p \le 2M-1 \le 2\left(\tfrac{N-6}{2}\right) - 1 = N-7 , so that N p 7 N-p \ge 7 . By induction, we deduce that N p N-p can be written as a sum of distinct primes, none of which are greater than m a x ( 11 , N p 7 ) \mathrm{max}(11,N-p-7) .

Note that p M + 1 > 11 p \ge M+1 > 11 and 2 p > 2 M N 6 > N 7 2p > 2M \ge N-6 > N-7 , so that p > N p 7 p > N-p-7 . Thus p > m a x ( 11 , N p 7 ) p > \mathrm{max}(11,N-p-7) , and hence we deduce that N N can be written as a sum of distinct integers, none of which are greater than p p . Since p N 7 m a x ( 11 , N 7 ) p \le N-7 \le \mathrm{max}(11,N-7) , we deduce that none of the primes that sum to N N are greater than m a x ( 11 , N 7 ) \mathrm{max}(11,N-7) , and so the desired result follows by induction.


Suppose that N 3 N \ge 3 is odd, and let p 1 , p 2 , . . . , p N p_1,p_2,...,p_N be the first N N primes. This list contains 2 2 and an even number of odd primes, so we deduce that 2 X = n = 1 N p n 2X \; =\; \sum_{n=1}^Np_n is an even integer (so that X N X \in \mathbb{N} ). Find an integer K K such that n = K N p n > X n = K + 1 N p n \sum_{n=K}^N p_n \; > \; X \; \ge \; \sum_{n=K+1}^Np_n If N = 3 N=3 then since 2 + 3 = 5 2+3=5 , we can split the first N = 3 N=3 primes into two sets with equal sums. If N = 5 N=5 we can split the first N = 5 N=5 prime numbers as 3 + 11 = 2 + 5 + 7 3+11 = 2 + 5 + 7 . Otherwise N 7 N \ge 7 , so that 2 n = 1 K p n 2 X = n = 1 N p n 2 + 3 + 5 + 7 + 11 + 13 + 17 = 58 2\sum_{n=1}^Kp_n \; \le \; 2X = \sum_{n=1}^Np_n \ge 2 + 3 +5 + 7 + 11+ 13 + 17 = 58 so that n = 1 K p n 29 \sum_{n=1}^Kp_n \ge 29 which implies that K 6 K \ge 6 . We can therefore suppose that K 6 K \ge 6 (so that p K 13 p_K \ge 13 ). There are now various cases to consider:

  • If n = K + 1 N p n = X \sum_{n=K+1}^N p_n = X , then we have a suitable split of the first N N primes.
  • If Z = X n = K + 1 N p n Z = X - \sum_{n=K+1}^N p_n is such that Z 7 Z \ge 7 , then we know that Z Z can be written as a sum of distinct primes, none of which exceed Z < p K Z < p_K . This means that we can write X X as the sum of a collection of a collection of distinct primes, all less than p K p_K , and the primes p K + 1 p_{K+1} to p N p_N .
  • if Z = 2 , 3 , 5 Z = 2,3,5 , then Z Z is prme and Z 5 < p K < p K + 1 Z \le 5 < p_K < p_{K+1} and X = Z + n = K + 1 N p n X = Z + \sum_{n=K+1}^Np_n .
  • If instead Z = 1 , 4 , 6 Z = 1,4,6 then X n = K + 2 N p n = Z + p K + 1 7 X - \sum_{n=K+2}^Np_n \; =\; Z + p_{K+1} \; \ge \; 7 since K 6 K \ge 6 , and so X n = K + 2 N p n X - \sum_{n=K+2}^Np_n can be written as a sum of distinct primes, none greater than m a x ( 11 , X n = K + 2 N p n 7 ) < p K + 1 \mathrm{max}\left(11,X - \sum_{n=K+2}^Np_n-7\right) < p_{K+1} (since p K + 1 > 11 p_{K+1} > 11 ), and so no prime is greater than p K p_K . Thus we have written X X as a sum of a subset of p 1 , p 2 , . . . , p N p_1,p_2,...,p_N

Is it always possible to divide the first 2 k + 1 2k+1 primes into two sets with equal sums? If not, what's the smallest k k that doesn't work? (It seems there's a way to do it if you assume the Goldbach conjecture, but that feels like a bit of a sledgehammer)

Chris Lewis - 1 year, 4 months ago

Log in to reply

Check out my extended proof.

Mark Hennings - 1 year, 3 months ago

How did you get the equality? I mean, is the same true for any odd value of n n ? Is there any analytical method of proving or disproving the statement that for any odd value of n n , the first n n prime numbers can always be grouped into two groups such that the sums of elements of the two groups are equal?

A Former Brilliant Member - 1 year, 3 months ago

I suppose the proposition should be: Any positive integer N≥7 can be written as the sum of distinct primes. Just "positive integers" is not enough.

S YW - 1 year, 3 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...