|
|
A323215
|
|
Numbers k such that row k of A322936 is not empty and has only primes as members.
|
|
1
|
|
|
|
OFFSET
|
1,1
|
|
COMMENTS
|
a is strongly prime to n if and only if a <= n is prime to n and a does not divide n-1. See the link to 'Strong Coprimality'. (Our terminology follows the plea of Knuth, Graham and Patashnik in Concrete Mathematics, p. 115.)
If there is at least one prime <= sqrt(n) that divides neither n nor n-1, then its square is strongly prime to n and not prime. If there does not exist such a prime, then the first Chebyshev function theta(sqrt(n)) = Sum_{p <= sqrt(n)} log(p) <= 2 log(n). Now it is known that theta(x) = x + O(x/log(x)), so this can't happen if n is sufficiently large. Thus the sequence is finite.
The largest n for which no such p exists appears to be 120. There are none between 121 and 10^7. It is possible that a sufficiently tight lower bound on theta together with a finite search can be used to prove that there are no other terms of the sequence. (End)
|
|
LINKS
|
|
|
MAPLE
|
filter:= proc(n) local k, found;
found:= false;
for k from 2 to n-2 do
if igcd(k, n)=1 and (n-1) mod k <> 0 then
found:= true;
if not isprime(k) then return false fi;
fi
od;
found
end proc:
|
|
MATHEMATICA
|
Select[Range[10^3], With[{n = #}, AllTrue[Select[Range[2, n], And[GCD[#, n] == 1, Mod[n - 1, #] != 0] &] /. {} -> {0}, PrimeQ]] &] (* Michael De Vlieger, Apr 01 2019 *)
|
|
PROG
|
(Sage) # uses[A322936row from A322936]
def isA323215(n):
return all(is_prime(p) for p in A322936row(n))
[n for n in (1..100) if isA323215(n)] # Peter Luschny, Apr 03 2019
|
|
CROSSREFS
|
|
|
KEYWORD
|
nonn,fini,full
|
|
AUTHOR
|
|
|
EXTENSIONS
|
|
|
STATUS
|
approved
|
|
|
|