Year-end appeal: Please make a donation to the OEIS Foundation to support ongoing development and maintenance of the OEIS. We are now in our 61st year, we have over 378,000 sequences, and we’ve reached 11,000 citations (which often say “discovered thanks to the OEIS”).
%I #70 Jul 10 2021 20:34:17
%S 2,6,9,15,21,24,30,34,36,49,51,54,64,75,76,84,90,91,114,115,117,120,
%T 121,132,135,141,154,156,174,175,184,187,201,205,210,216,217,220,231,
%U 244,246,252,285,286,294,297,300,301,304,321,322,324,327,330,339,360,364
%N a(n) is the smallest number j such that if x >= j there are at least n primes between x and 2x exclusively.
%C For all m >= a(n) there are at least n primes between m and 2m exclusively. This calculation relies on the fact that pi(2m) - pi(m) > m/(3*log(m)) for m >= 5. This is one more than the terms of A084139 with offset changed from 0 to 1.
%C For n > 5889, pi(2n) - pi(n) > f(2, 2n) - f(3, n) where f(k, x) = x/log x * (1 + 1/log x + k/(log x)^2). This may be useful for checking larger terms. The constant 3 can be improved at the cost of an increase in the constant 5889. - _Charles R Greathouse IV_, May 02 2012
%C A168421(n) = nextprime(a(n)), where nextprime(x) is the next prime >= x. - _John W. Nicholson_, Dec 21 2012
%C a(1) = ceiling((A104272(1)+1)/2) modifies the only even prime, 2; which has been stated, in Formula, as a(1) = A104272(1); for all others, a(n) = (A104272(n)+1)/2 = ceiling ((A104272(n)+1)/2). - _John W. Nicholson_, Dec 24 2012
%C Srinivasan's Lemma (2014): previousprime(a(n)) = p_(k-n) < (p_k)/2, where the n-th Ramanujan Prime R_n is the k-th prime p_k, and with n > 1. Proof: By the minimality of R_n, the interval ((p_k)/2,p_k] contains exactly n primes, so p_(k-n) < (p_k)/2. - Copied and adapted from a comment by _Jonathan Sondow_ in A168421 by _John W. Nicholson_, Feb 17 2015
%D Paulo Ribenboim, The Little Book of Big Primes, Springer-Verlag, 1991, p. 140.
%D Paulo Ribenboim, The Little Book of Bigger Primes, Springer-Verlag, 2004, p. 181.
%H Vladimir Shevelev, Charles R. Greathouse IV, Peter J. C. Moses, <a href="https://cs.uwaterloo.ca/journals/JIS/VOL16/Moses/moses1.html">On intervals (kn, (k+1)n) containing a prime for all n>1</a>, Journal of Integer Sequences, Vol. 16 (2013), Article 13.7.3. <a href="http://arxiv.org/abs/1212.2785">arXiv version</a>, arXiv:1212.2785 [math.NT], 2012.
%H Anitha Srinivasan, <a href="http://www.emis.de/journals/INTEGERS/papers/o19/o19.Abstract.html">An upper bound for Ramanujan primes</a>, Integers, 19 (2014), #A19.
%H Eric Weisstein's World of Mathematics, <a href="http://mathworld.wolfram.com/BertrandsPostulate.html">Bertrand's Postulate</a>.
%F a(1) = A104272(1); for n >= 2, a(n) = (A104272(n)+1)/2. - _Vladimir Shevelev_, Dec 07 2012
%F a(n) = ceiling((A104272(n)+1)/2) for n >= 1. - _John W. Nicholson_, Dec 24 2012
%e a(11)=51 since there are at least 11 primes between m and 2m for all m >= 51 and this is not true for any m < 51. Although a(100)=720 is not listed, for all m >= 720, there are at least 100 primes between m and 2m.
%t nn = 60;
%t R = Table[0, {nn}]; s = 0;
%t Do[If[PrimeQ[k], s++]; If[PrimeQ[k/2], s--]; If[s < nn, R[[s + 1]] = k], {k, Prime[3 nn]}];
%t A104272 = R + 1;
%t Ceiling[(A104272 + 1)/2] (* _Jean-François Alcover_, Nov 07 2018, after _T. D. Noe_ in A104272 *)
%Y Cf. A108954, A060715, A060756, A084138, A084139, A084141, A084142, A168421, A168425.
%K nonn
%O 1,1
%A _Harry J. Smith_, May 15 2003