[seqfan] Re: All integers a(n) > 20 are in sequence A246778
Vladimir Shevelev
shevelev at bgu.ac.il
Sat Nov 28 14:29:24 CET 2015
On the one hand, your result does more
plausible the truth
of Firoozbakht's (1982)-conjecture. As you
proved, from this conjecture it follows that,
for all k>9,
p_(k+1) - p_k < log^2(p_k) - log (p_k) -1. (1)
On the other hand, A. Granvill (1995) in his
improved Cramer's model obtained that
lim sup (p_(k+1) - p_k)/log^2(p_k) >=
2*e^(-c)=1.1229...,
which contradicts (1). Cf.
http://www.dartmouth.edu/~chance/chance
_news/for_chance_news/Riemann/cramer.pdf_
(Scand. Actuarial J., 1 (1995), 12-28).
Also, of course, Granvill's result cannot
contradict your last theorem for A246778.
Best regards,
Vladimir
_______________________________________
From: SeqFan [seqfan-bounces at list.seqfan.eu] on behalf of Alexei Kourbatov [akourbatov at gmail.com]
Sent: 28 November 2015 03:41
To: Sequence Fanatics Discussion list
Subject: [seqfan] All integers a(n) > 20 are in sequence A246778
Dear Seqfans,
Just added a new theorem in A246778 showing that all terms a(n)>20 are in
the sequence (just as conjectured by Farideh Firoozbakht) - provided that
prime gaps near x are less than x^0.75. Any feedback will be greatly
appreciated. (If you happen to have a nicer/shorter proof, that would be
great!)
Alexei
_______________________________________________
Seqfan Mailing list - http://list.seqfan.eu/
More information about the SeqFan
mailing list