|
|
A220293
|
|
Chebyshev numbers C_2(n): a(n) is the smallest number such that if x >= a(n), then theta(x) - theta(x/2) >= n*log(x), where theta(x) = sum_{prime p <= x} log p.
|
|
4
|
|
|
11, 17, 29, 41, 47, 59, 67, 71, 97, 101, 107, 127, 149, 151, 167, 179, 223, 229, 233, 239, 241, 263, 269, 281, 307, 311, 347, 349, 367, 373, 401, 409, 419, 431, 433, 443, 461, 487, 503, 569, 571, 587, 593, 599, 601, 607, 641, 643, 647, 653
(list;
graph;
refs;
listen;
history;
text;
internal format)
|
|
|
OFFSET
|
1,1
|
|
COMMENTS
|
Up to a(100)=1489, only two terms of the sequence (a(17)=223 and a(36)=443) are not Ramanujan numbers (A104272), and the sequence is missing only the following Ramanujan numbers up to 1489: 2, 181, 227, 439, 491, 1283, and 1301. The latter observation shows how closely the ratio theta(x)/log(x) approximates the number of primes <= x (i.e., pi(x)).
A generalization: for a real number v>1, the v-Chebyshev number C_v(n) is the smallest integer k such that if x>=k, then theta(x)-theta(x/v)>=n*log x. In particular, a(n)=C_2(n). For another example, if v=4/3, then, at least up to 3319, all (4/3)-Chebyshev numbers are (4/3)-Ramanujan primes as in Shevelev's link (cf. A193880, where c=1/v=3/4 is excepted), and in this case the sequence is missing only the following (4/3)-Ramanujan numbers up to 3319: 11 and 1567.
Like Chebyshev numbers, all v-Chebyshev numbers are primes.
|
|
LINKS
|
N. Amersi, O. Beckwith, S. J. Miller, R. Ronan, J. Sondow, Generalized Ramanujan primes, Combinatorial and Additive Number Theory, Springer Proc. in Math. & Stat., CANT 2011 and 2012, Vol. 101 (2014), 1-13
|
|
FORMULA
|
For n >= 2, A104272(n) <= a(n-1) <= prime(3n).
|
|
CROSSREFS
|
|
|
KEYWORD
|
nonn
|
|
AUTHOR
|
|
|
STATUS
|
approved
|
|
|
|