|
%I #23 Jan 03 2020 05:32:21
%S 2,3,7,12,15,17,20,21,25,26,27,32,34,35,36,39,40,42,51,52,56,57,58,62,
%T 63,64,65,69,70,72,76,77,80,81,82,86,87,88,90,91,92,93,94,95,99,100,
%U 104,105,106,110,111,112,115,116,117,122,125,128,130,133,134,135,140,141,142,145,146,147,148,154,155,156,158,159,160,161,162,164,165,168,169,171,172,175,176,177,178
%N Positive integers not occurring in A330614 (subsequence of first differences with indices of primes gives back the original).
%C Sequence A330614 has an interesting "fractal" structure of interleaved subsequences. It is computed in a nearly greedy manner as the lexicographically smallest sequence of distinct terms which has a certain property. One might expect that this yields a permutation of the positive integers, but it turns out that some numbers will never occur. These are listed here, in the hope that more can be found out about this subset.
%C While the first terms are isolated, the sequence appears to have a density of > 3/5 already around n = 500, 2/3 around n = 1500, and almost 3/4 near n = 2*10^4. Is this the limiting density, or is it asymptotically increasing to a larger value?
%H M. F. Hasler, <a href="/A330593/b330593.txt">Table of n, a(n) for n = 1..18900</a>: terms that can be deduced from A330614(1..10^4).
%H M. F. Hasler, <a href="/A330593/a330593.gif">Plot of the density averaged over sqrt(a(n)) terms</a>, Dec 30 2019.
%o (PARI) A=setminus([1..2*#A330614],Set(A330614)) \\ with a sufficiently large vector A330614.
%Y Complement of the range of A330614 in the positive integers A000027.
%K nonn
%O 1,1
%A _M. F. Hasler_, Dec 27 2019
|
