Welcome to our community

Be a part of something great, join today!

Prove/disprove inequality involving sums of consecutive twin prime pairs - - (My own problem)

checkittwice

Member
Apr 3, 2012
37
.
.

[tex]Let \ \ p_n \ \ = \ \ the \ \ nth \ \ prime \ \ number.[/tex]


Examples:


[tex]p_1 \ = \ 2[/tex]

[tex]p_2 \ = \ 3[/tex]

[tex]p_3 \ = \ 5[/tex]

[tex]p_4 \ = \ 7[/tex]


- - - - - - - - - - - - - - - - - - - - - - - - - - - -


[tex]Let \ \ n \ \ belong \ \ to \ \ the \ \ set \ \ of \ \ positive \ \ integers.[/tex]



Prove (or disprove) the following:


[tex]p_n \ + \ p_{n + 1} \ \ \ge \ \ p_{n + 2} \ + \ p_{n - 2}, \ \ \ for \ \ all \ \ n \ \ge \ 4.[/tex]





Examples:


[tex] \ \ 7 \ + \ 11 \ \ > \ \ 13 \ + \ \ 3[/tex]

[tex]19 \ + \ 23 \ \ = \ \ 29 \ + \ 13[/tex]
 
Last edited:

CaptainBlack

Well-known member
Jan 26, 2012
890
.
.

[tex]Let \ \ p_n \ \ = \ \ the \ \ nth \ \ prime \ \ number.[/tex]


Examples:


[tex]p_1 \ = \ 2[/tex]

[tex]p_2 \ = \ 3[/tex]

[tex]p_3 \ = \ 5[/tex]

[tex]p_4 \ = \ 7[/tex]


- - - - - - - - - - - - - - - - - - - - - - - - - - - -


[tex]Let \ \ n \ \ belong \ \ to \ \ the \ \ set \ \ of \ \ positive \ \ integers.[/tex]



Prove (or disprove) the following:


[tex]p_n \ + \ p_{n + 1} \ \ \ge \ \ p_{n + 2} \ + \ p_{n - 2}, \ \ \ for \ \ all \ \ n \ \ge \ 4.[/tex][/tex]





Examples:


[tex] \ \ 7 \ + \ 11 \ \ > \ \ 13 \ + \ \ 3[/tex]

[tex]19 \ + \ 23 \ \ = \ \ 29 \ + \ 13[/tex]
Asymtotically \(p_n+p_{n+1} \sim (2n+1)\log(x)\) and \(p_{n-2}+p_{n+2} \sim 2n\log(n)\).

so the inequality eventually holds, and how many terms we need to check explicitly before we can rely on the asymtotics can probably be determined from (on second thoughts it can't, the inequalities have too wide a spread):

\(n\log(n)+n\log(\log(n))-n<p_n<n\log(n)+n\log(\log(n)), \ \ n\ge 6\)

It appears to fail for \(n=29\), when \(p_n=109,\ p_{n+1}=113,\ p_{n-2}=103,\ p_{n+2}=127\)

There are also plenty of other counter examples for primes less than \(10^6\).

So, the new question is: Is there a \(n_0\) such that for all \(n>n_0\) the inequality holds? I would guess the answer is no.

CB
 
Last edited: