|
|
|
|
3, 7, 8, 31, 127, 8191, 131071, 524287, 2147483647, 2305843009213693951, 618970019642690137449562111, 162259276829213363391578010288127, 170141183460469231731687303715884105727
(list;
graph;
refs;
listen;
history;
text;
internal format)
|
|
|
OFFSET
|
1,1
|
|
COMMENTS
|
Likely the union of the Mersenne primes A000668 and the {8}. [Presence of Mersenne primes M is obvious because A085392(M+1)=2 and A085392(M)=1.
Absence of other primes p is also clear because A085392(p)=1 and A085392(p+1) >=3 because it contains at least one odd prime factor if not of the Mersenne type.
For composite candidates c, we search adjacent c+1 and c with largest noncomposite divisors 2 and 1 or 3 and 2. The first branch enforces c=2 which is immediately discarded. The second branch searches for a power of 3 adjacent to a power of 2, and the solution to this exponential diophantine equation 3^x-2^y=1 is believed to lead only to the 8 (see the Weger review in the link).]
|
|
LINKS
|
B. M. M. de Weger, Book review, Bull. Am. Math. Soc. 25 (1991) 145-146.
|
|
EXAMPLE
|
Number 8 is in sequence because the difference between A085392(9)=3 and A085392(8)=2 is 1.
31 is in sequence because the difference between A085392(32)=2 and A085392(31)=1 is 1.
|
|
PROG
|
(PARI) gpd(n) = if (n==1, 1, n/factor(n)[1, 1]);
gpf(n) = if (n==1, 1, vecmax(factor(n)[, 1]));
f(n) = gpf(gpd(n));
|
|
CROSSREFS
|
|
|
KEYWORD
|
nonn
|
|
AUTHOR
|
|
|
STATUS
|
approved
|
|
|
|