Firoozbakht's conjecture and maximal gaps

504 Views Asked by At

In the Wikipedia article, it seems to me as if it's implied that it is enough to check the conjecture only for maximal gaps (numbers $n$ s.t. $\forall k<n:g_n>g_k$).

I.e it holds that $P_{k+1}<P_k^{1+\frac{1}{k}}\Rightarrow P_{m+1}<P_m^{1+\frac{1}{m}}$ where $P_{k+1}-P_{k}$ is a maximal gap and $m<k$. (Or some similar statement that involves maximal gaps.)

Is this really the case or perhaps I misread the Wikipedia article? I tried to prove it using high school manipulations, but without much success.

2

There are 2 best solutions below

2
On

Define $q(x)$ to be the smallest prime greater than $x$ and $$ f(p)=\begin{cases}1&\text{, }p\text{ is prime and }q(p)<p^{1+1/\pi(p)}\\ 0&\text{, otherwise}\end{cases} $$

Then Firoozbakht's conjecture is that $f(p)=1$ for all primes $p$. You seem to be asking if $f(p)=1$ for all primes $p$ beginning a maximal gap implies Firoozbakht's conjecture, and it does.

I should also mention that Firoozbakht's conjecture is believed to be false. It is expected that for any $k<2/e^\gamma=1.1229\ldots$ there are infinitely many prime gaps of length greater than $kp\log^2p$ and Firoozbakht's conjecture fails if there is even one with $k\ge1.$ It was a reasonable conjecture in 1982 when it was first posed but advances in the mid-80s and later have convinced number theorists that it's highly unlikely.

1
On

The implication "if a certain sufficient condition $(*)$ for Firoozbakht's conjecture is true for a maximal gap starting at a prime $p_k$, $k>9$, then it is also true for all prime gaps between this maximal gap and the next one" is shown in arXiv:1506.03042 (Journal of Integer Sequences, 18, 2015, article 15.11.2) Upper bounds for prime gaps related to Firoozbakht's conjecture; see Section 4, Theorem 3 (page 4) and Remark (i) (page 5). Namely, the implication is based on the following sufficient condition for Firoozbakht's conjecture (Theorem 3): if $$p_{k+1}-p_k<\log^2 p_k - \log p_k - 1.17 \quad (*)$$ for $k>9$, then $p_{k+1}<p_k^{1+1/k}$. Just observe that $\log^2 p_k - \log p_k - 1.17$ is an increasing function of $p_k$ and use the definition of maximal gaps.