-3
$\begingroup$

A semiprime $s$ is a positive integer that is the product of two prime numbers, see Semiprine the encyclopedia Wikipedia. A well-known inequality, with applications, that involves prime numbers is the named Bonse's inequality, I add as reference from the Wikipedia Bonse's inequality.

I wondered if it is possible to create a similar inequality where $s_k$ denotes the $k$-th semiprime (thus the sequence A001358 from the OEIS)

$$m(n)\cdot (s_{n+1})^{a}<\left(\prod_{k=1}^n s_k\right)^b$$ for a suitable arithmetic function (or sequence) $m(n)$ and constants $a$ and $b$ (I've added this function $m(n)$ and constant $b$ with the purpose to provide flexibility in the research of the inequality that we evoke).

Question. How to get a sharper inequality* involving semiprimes $s_k$ $$(s_{n+1})^{a}<\frac{1}{m(n)}\left(\prod_{k=1}^n s_k\right)^b\tag{1}$$ that holds $\forall n>N$ for a suitable choice of $N$, and for a suitable choice of a function $m(n)>0$ and constants $a$ and $b$? Many thanks

*If in your investigations you get a remarkable inequality, or asymptotic inequality, that does not fit exactly to the previous type of inequality $(1)$ I think that it is reasonable that you can to feel free to add it as an answer because I am asking for a good version of a Bonse's inequality for semiprimes.

With the words a sharper inequality I mean that your inequality of the type $(1)$ have good mathematical content/meaning, that it is a good Bonse's inequality for semiprimes. To emphasize we take yours $m(n)$ as a positive arithmetic function, thus $m(n)>0$ for all $n>N$.

$\endgroup$
3
  • $\begingroup$ In the linked Wikipedia is cited the bibliography for Bonse's inequality. An interesting article, that also refers Bonse's inequality, is by Robert J. Betts, Using Bonse's Inequality for Find Upper Bounds on Prime Gaps, Journal of Integer Sequences, Vol. 10 (2007) Article 07.3.8 $\endgroup$
    – user142929
    Commented Sep 29, 2019 at 10:41
  • 6
    $\begingroup$ For the record, Bonse's inequality is far from sharp. $p_1\dots p_n$ is on the order of $e^{p_n}$, while $p_{n+1}$ is barely larger than $p_n$ (e.g. it's smaller than $2p_n$). You can get pretty precise bounds using this and this $\endgroup$
    – Wojowu
    Commented Sep 29, 2019 at 10:41
  • $\begingroup$ Yes I accept your critic, I am looking for a good version of Bonse's inequality for semiprimes. Many thanks @Wojowu $\endgroup$
    – user142929
    Commented Sep 29, 2019 at 10:43

1 Answer 1

2
$\begingroup$

You have that k choose 2 semi primes (and more) are less than $(p_k)^2$ which in turn from Bonse's inequality is less than the primorial involving $p_{k-1}$ which is the product of about k/2 semiprimes.
You may be able to arrange the product so that the semiprime factors have small indices.

Gerhard "Offering A Semi-Baked Prime Idea" Paseman, 2020.02.04.

$\endgroup$
1
  • $\begingroup$ Many thanks for your attention and good apporach. Due my low reputation when I've casted an upvoted, it doesnt' work (my upvote is just positive feedback, and it doesn't show publicy). $\endgroup$
    – user142929
    Commented Feb 4, 2020 at 20:22

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .