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 #26 Aug 08 2019 01:57:55
%S 31,89,127,223,233,431,601,881,911,1103,1801,2089,2351,3191,3391,4513,
%T 5209,6361,8191,9623,9719,11447,11471,13367,14951,15193,15809,18041,
%U 18121,18199,18287,20231,23279,23671,39551,43441,50023,53993,54217,55441,55871,59233
%N Odd primes which can never divide 2^a+2^b+1.
%C Contains the Mersenne primes M_p for p>3 as a subsequence, as 2^a+2^b cannot exceed 2^(p-1)+2^(p-2) which is less than 2^p-2 is p>3.
%C Mariusz Skałba conjectures that this sequence has density zero among all primes but contains infinitely many primes based on the following observations. For any prime p in this sequence, the multiplicative order of 2 modulo p is <p^0.8 (Erdős conjectures that the set of such primes must have density zero among all primes). Moreover, any number of the form 2^m-1 the number of whose prime factors counted with multiplicity is <log m/log 3 has at least one prime factor in this sequence (the infinitude of such numbers may be more tractable than the infinitude of Mersenne primes). - _Tomohiro Yamada_, Aug 08 2019
%H Robert Israel, <a href="/A179113/b179113.txt">Table of n, a(n) for n = 1..129</a>
%H Mariusz Skałba, <a href="https://doi.org/10.1007/s00017-004-0211-x">Two conjectures on primes dividing 2^a+2^b+1</a>, Elemente der Mathematik 59 (2004), issue 4, pp. 171-173.
%e 31 is on the list as you can't sum any two of {1, 2, 4, 8, 16} to make 30 (mod 31).
%p N:= 10000; # to test the first N primes for membership
%p A179113:= proc(p)
%p local x, R;
%p x:= 1; R:= {};
%p do
%p R:= R union {p-1-x};
%p if member(x,R) then return(false) end if;
%p x:= 2*x mod p;
%p if x = 1 then return(true) end if;
%p end do;
%p end proc;
%p select(A179113,[seq(ithprime(i),i=2..N)]);
%p # _Robert Israel_, May 19 2013
%t n = 10000; (* to test the first n primes for membership *) A179113[p_] := Module[{x = 1, r = {}}, While[True, r = r ~Union~ {p-1-x}; If[MemberQ[r, x], Return[False]]; x = Mod[2*x, p]; If[x == 1, Return[True]]]];Reap[Do[If[A179113[p], Print[p]; Sow[p]], {p, Prime /@ Range[2, n]}]][[2, 1]] (* _Jean-François Alcover_, Dec 02 2013, translated from _Robert Israel_'s Maple program *)
%o (PARI) forprime(p=3,1000,pol=x+O(x^p);t=2;while(t-1,pol+=x^t;t=t*2%p);pol2=pol*pol;if(!polcoeff(pol2,p-1),print1(p", ")))
%K nonn
%O 1,1
%A _Phil Carmody_, Jan 04 2011
%E More terms from _Robert Israel_, May 19 2013