Are there any sequences of positive integers $ 1 \leq a_{1} < a_{2} < a_{3} < \ldots$ such that for each integer $ n$, the set $ \left\{a_{k} + n\ |\ k = 1, 2, 3, \ldots\right\}$ contains finitely many prime numbers?
Problem
Source: Romanian TST 2 2008, Problem 2
Tags: number theory, prime numbers, number theory proposed
07.06.2008 20:17
finitly many primes? is it mean that the number of primes is defined? For example does sequence $ a_k = k!$ satisfy the conditions?
07.06.2008 20:28
It's an open problem if $ k!$ works (it's not known if $ k!+1$ is prime finitely often or not). But one can take $ a_k$ as a multiple of $ k!$ such that $ a_k \pm 1$ are not prime.
07.06.2008 20:32
OK, then we take $ a_k=k!+1$
07.06.2008 20:33
No, as then for $ n=0$ you have no proof of finiteness of primes in $ a_k+n$.
07.06.2008 20:33
Let $ a_k = k! + 2$, then $ a_k + n$ had only finetely prime numbers for any $ n\in N$. I am sorry $ n\in Z$. Therefore we can take \[ a_k=(k!)^6\], then $ a_k-1=((k!)^3-1)((k!)^3+1$ and $ a_k+1=((k!)^2+1)(k!^4-k!^2+1).$
07.06.2008 20:39
May be I misunderstand the condition, but we must prove, that for every integer $ n$ the set $ {a_k+n|k=1,2,3,...}$ contains finitely number of primes. Does we vary the $ k$ or $ n$?
07.06.2008 20:56
The set is explicitely given by holding $ n$ constant and varying $ k$.
07.06.2008 20:58
if the $ n$ constant, then what is the problem with $ k!+1$? it's at most one prime. Or we consider unions of sets for all n?
07.06.2008 23:07
For all $ n$ the set corresponding to $ n$ shall have only finitely many primes.
08.06.2008 14:57
You can use also $ a_k=4(k!)^4$ and for n=1 use Sophie Germain while for n=-1 use difference of squares
11.06.2008 14:34
Take ak=(k!)^3