Define $a_1=2$ and

$$ a_n = \left\{ \begin{array}{ll} p_na_{n-1}\quad\quad n\quad \text{odd}\\ p_n+a_{n-1}\quad n\quad \text{even}\\ \end{array} \right. $$

where $p_n$ is the nth prime number. The first few values are $2,5,25,32,352,365,\cdots$.

I've done some research and found that for $n<25$, this sequence produces only 2 primes in this range, $a_1=2$ and $a_2=5$. Is there a methodology for going about showing whether or not the sequence actually produces no more primes or not? I realize that $a_{4k-2}$ for $k\in\mathbb{N}$ are the only contenders since the others are either even and/or divisible by $p_i$. The numbers get very large, so it becomes difficult to check higher values of $n$.

Any thoughts appreciated. Thanks

  • 1
    $\begingroup$ You've got the odd/even operation flipped. $\endgroup$ – Samuel Feb 24 '18 at 15:53
  • $\begingroup$ Yes, but 2 is first. I'll add the first values to the question. $\endgroup$ – Samuel Feb 24 '18 at 15:57
  • $\begingroup$ Those first few values suggest the sequence alternates even numbers with multiples of 5. You have already found values that break that pattern, right? $\endgroup$ – Robert Soupe Feb 24 '18 at 16:50
  • 1
    $\begingroup$ Correct, that pattern doesn't continue. $\endgroup$ – Samuel Feb 24 '18 at 17:12

According to my calculation in PARI/GP $$a_{518}$$ is a probable prime with $792$ digits and $$a_{1226}$$ is a probable prime with $2143$ digits.

  • $\begingroup$ Thanks for inspecting. Interesting. What algorithm are you using? $\endgroup$ – Samuel Feb 24 '18 at 16:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.