4

Is this true that $n\log\left(\frac{p_n}{p_{n+1}}\right)$ is bounded, where $p_n$ is the $n$-th prime number?

Wei Wang
  • 458
Papagon
  • 315

2 Answers2

4

Seems unbounded:

Let $g_n = p_{n+1} - p_n$ be the prime gap, then Westzynthius's result (see link below) states that $\lim\sup \left[ g_n/(\log p_n) \right] = \infty$, hence

$$\lim \sup n \log(p_{n+1}/p_n) = \lim \sup n \log (1 + g_n/p_n) = \lim \sup n g_n/ p _n = \lim \sup g_n/\log n = \infty$$

http://en.wikipedia.org/wiki/Cram%C3%A9r%27s_conjecture

PA6OTA
  • 1,972
  • 10
  • 12
-2

$\frac{p_n}{p_{n+1}}<1$ for all $n$, so when $n\to\infty$, $(\frac{p_n}{p_{n+1}})^n\to0$. so $$n\log(\frac{p_n}{p_{n+1}})=\log(\frac{p_n}{p_{n+1}})^n\to-\infty$$ so this is not bounded

Martial
  • 1,694
  • -1 because $\lim (p_n/p_{n+1})^n \ne 0$. Probably does not exist – PA6OTA May 16 '14 at 15:05
  • 1
    And also because this is false reasoning... if $x_n \in [0,1[$ satisfies $x_n \to x \in [0,1[$, then $y_n = \sqrt[n]{x_n}$ satisfies $0 \le y_n < 1$ and $y_n^n \to x$. So this reasoning is not sufficient to show $(p_n/p_{n+1})^n \to 0$. – Patrick Da Silva May 16 '14 at 15:05