Hi everyone,
Hope all is well. I would like to propose the following sequence and I would really appreciate it if you could tell me if it’s suitable for the OEIS.
For a fixed positive integer n, compute the largest prime factor of k raised to the nth power plus 1 for k = 1, 2, 3, ...
The sequence term a(n) is the first value of k for which this largest prime factor is less than or equal to the previous one.
For a fixed n >= 1, let
![]()
where gpf(m) denotes the largest prime factor of m.
Starting with k = 1, compute the sequence
![]()
and continue as long as these
values are strictly increasing. Let
be the first value of
for which the increase
fails, i.e.,
![]()
Equivalently, a(n)-1 is the length of the initial strictly increasing run of largest prime factors of k^n+1.
The first terms are
![]()
Examples:
![]()
Since
, the increase fails at
, hence
.
![]()
coming from
![]()
Since 7 <13 , the increase fails at k=5, hence a(3) = 5.
![]()
coming from
![]()
Since 1201 < 1297 the increase fails at k = 7, hence a(4) = 7.
First terms:
![]()
Conjecture. For every nonconstant integer polynomial whose values are eventually greater than 1, the largest prime factor of the polynomial values has infinitely many descents.
A similar sequence (and conjecture) could be obtained from n^k+1.
Best,
Ali