7
$\begingroup$

Is there any set $X$ which is a density 0 subset of $N^*$ and we already know that there are infinitely many primes in it, beside examples which come from $x^2+y^4$(or its proof)?

Problem1: In particular, is it already proved that there exist $c>1$, s.t. $A_c=\{n\in\mathbb{N}^*| \exists k\in\mathbb{N}^* , n=[k^c]\}$ contains infinitely many primes?

If this problem can not be solved by existing methods, can the following one be solved by existing methods?

Problem2: for all $c>1$, $\exists d(c), c>d(c)>0$, s.t. $A_c=\{n\in\mathbb{N}^*| \exists k\in\mathbb{N}^* , n=[k^{e}], e\in (c-d(c),c+d(c))\}$ contains infinitely many primes.

This is only a naive question, thanks in advance.

Improve this question
$\endgroup$
1
  • 3
    $\begingroup$ As stated this has trivial answers—for example, the set $X=\{$all prime numbers$\}$ has the required properties. You might want to be more specific about the types of sets you intend to consider. $\endgroup$
    – Greg Martin
    Dec 1, 2020 at 6:59

2 Answers 2

Reset to default
10
$\begingroup$

Yes, there is a $c > 1$ for which infinitely many numbers of the form $\lfloor k^{c} \rfloor$ are prime. The first result of this type was proven in Ilya Piatetski-Shapiro's Ph.D. thesis (written in 1954 under the direction of Alexander Buchstab) and holds for any $1 \leq c \leq 12/11$. A reference (from Wikipedia) is Pyateckiĭ-Šapiro, I.I. (1953). "On the distribution of prime numbers in sequences of the form [f(n)]". Mat. Sbornik N.S. 33 (75): 559–566.

The largest currently know range of $c$ values is $1 \leq c < 243/205$ due to Rivat and Wu. (Glasgow Math Journal, 2001, volume 43, no. 2, 237-254.)

Improve this answer
$\endgroup$
3
  • 2
    $\begingroup$ Do experts believe that it holds for all real $c$ from $1$ to $2$? And what is believed to be the maximal interval for which it holds? $\endgroup$
    – user21820
    Dec 1, 2020 at 6:50
  • 1
    $\begingroup$ Yes. In a JNT article from 1983, Heath-Brown states that $1 < c < 2$ one expects $|A_{c} \cap [1,x]| \sim x/(c \log(x))$ as $x \to \infty$. Of course, there are no prime numbers of the form $\lfloor n^{2} \rfloor$, but it is natural to guess that if $c > 1$ is not an integer, then there are infinitely many primes of the form $\lfloor n^{c} \rfloor$. Indeed, Deshouillers has proven that the Lebesgue measure of the set of $c$ for which there are finitely many primes in $A_{c}$ is zero. $\endgroup$
    – Jeremy Rouse
    Dec 1, 2020 at 13:37
  • $\begingroup$ Oh okay thanks! I'm surprised that despite Deshouillers' result, we still don't know about the $n^2{+}1$ conjecture. $\endgroup$
    – user21820
    Dec 1, 2020 at 19:50
6
$\begingroup$

I should mention that aside from primes represented by the polynomial $x^2 + y^4$, we may further thin the sequence by insisting that $y$ is also prime; this is a result of Heath-Brown and Li. Moreover, in 2001 Heath-Brown showed that the binary cubic form $x^3 + 2y^3$ represents infinitely many primes, thereby resolving an old question of Hardy which asked whether there are infinitely many primes which are sums of three cubes. Heath-Brown's result was subsequently generalized in joint work with B. Moroz in 2002.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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