OFFSET
1,2
COMMENTS
Every prime is in the sequence. If n==0 (mod 4), then it is not in the sequence. Moreover, if, for a prime p, n==0 (mod p^p), then n is not in the sequence. Indeed, if n=(p^p)*k, then n'=(p^p)'*k+p^p*k'=p^p(k+k')>=n, analogously, n''>=n', etc.
Conjecture. For every sufficiently large greater q of twin primes, the sequence contains infinite increasing sequence {s_n} of semiprimes beginning with 2*(q-2), such that (s_n)'=s_(n-1).
This conjecture is true, if 1) there exist infinitely many twin primes; 2) there exists n_0, such that for every prime p>n_0, number 2*p is sum of two primes r,t, for which r*t-2 is prime.
Proof. Let q>=n_0. Put s_1=2(q-2). By the condition, 2(q-2)=r+t, such that r*t-2 is prime. Put s_2=r*t and s_3=2(r*t-2). Then (s_3)'=2'*(r*t-2)+2*(r*t-2)'=s_2; (s_2)'=r+t=s_1. Continuing this process, we get an infinite sequence of semiprimes and every semiprime is a number whose sequence of number-derivatives is monotonically decreasing.
LINKS
Alois P. Heinz, Table of n, a(n) for n = 1..1000
EXAMPLE
MAPLE
d:= n-> n*add(i[2]/i[1], i=ifactors(n)[2]):
a:= proc(n) option remember; local i, j, k;
for k from 1 +`if`(n=1, 0, a(n-1)) do
i, j:= d(k), k;
while i<>0 and i<j do i, j:= d(i), i od;
if i=0 then return k fi
od
end:
seq(a(n), n=1..80); # Alois P. Heinz, Jul 22 2011
MATHEMATICA
d[n_] := If[n < 2, 0, n*Sum[i[[2]]/i[[1]], {i, FactorInteger[n]}]];
a[n_] := a[n] = For[k = 1+If[n == 1, 0, a[n-1]], True, k++, {i, j} = {d[k], k}; While[i != 0 && i < j, {i, j} = {d[i], i}]; If [i == 0, Return[k]]];
Array[a, 80] (* Jean-François Alcover, Apr 06 2017, after Alois P. Heinz *)
CROSSREFS
KEYWORD
nonn
AUTHOR
Vladimir Shevelev, Jun 25 2011
STATUS
approved