Almost-prime values of polynomials at prime arguments

A. J. Irving
DOI: https://doi.org/10.1112/blms/bdv035
2014-10-13
Abstract:We consider almost-primes of the form $f(p)$ where $f$ is an irreducible polynomial over $\mathbb Z$ and $p$ runs over primes. We improve a result of Richert for polynomials of degree at least $3$. In particular we show that, when the degree is large, there are infinitely many primes $p$ for which $f(p)$ has at most $°f+O(\log°f)$ prime factors.
Number Theory
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the distribution of almost - primes in the values of specific polynomials. Specifically, A.J. Irving, the author, focuses on almost - prime values of the form \( f(p) \), where \( f \) is an irreducible polynomial and \( p \) is a prime number. ### Problem Background In number theory, a classic problem is to prove that if the values \( f(n) \) of an irreducible polynomial \( f\in\mathbb{Z}[x] \) have no fixed prime factors, then there exist infinitely many \( n \) such that \( f(n) \) is prime. This problem has been solved for linear polynomials by Dirichlet's theorem, but it remains an open problem for higher - degree polynomials. By sieve methods, it can be proved that there exist infinitely many \( n \) such that \( f(n) \) has a finite number of prime factors. Previous results by Richert showed that for polynomials \( f \) of degree at least 3, there exist infinitely many \( n \) such that \( f(n) \) is the product of at most \( \deg f+ 1 \) prime factors (i.e., \( P_{\deg f + 1} \)). Iwaniec improved the case for the quadratic polynomial \( n^{2}+1 \), proving it to be \( P_{2} \). ### Main Contributions of the Paper Irving mainly improved Richert's results in this paper, especially for high - degree polynomials \( f(p) \) (where \( p \) is a prime number). He proved that when the degree of the polynomial is large, there exist infinitely many prime numbers \( p \) such that \( f(p) \) has at most \( \deg f+O(\log\deg f) \) prime factors. This result is close to Richert's result for \( f(n) \), but the factor 2 is removed and a lower - order term \( O(\log\deg f) \) is introduced. ### Main Conclusions The core conclusions of the paper are: - For any given irreducible polynomial \( f\in\mathbb{Z}[x] \), assuming it has no fixed prime factors and has a positive leading coefficient, there exist infinitely many prime numbers \( p \) such that \( f(p) \) is the product of at most \( r \) prime factors, where \( r \) is related to the degree \( \deg f \) of the polynomial. - When the degree of the polynomial is large enough, \( r \) can be expressed as \( \deg f + c\log\deg f+O(1) \), where \( c = 3.120... \) ### Methods and Techniques To obtain the above results, the author used the weighted sieve method and combined it with the two - dimensional sieve method to estimate certain terms more effectively. In addition, he also utilized the Bombieri - Vinogradov theorem and related results of the Elliott - Halberstam conjecture to optimize the application range of the sieve method. In conclusion, this paper significantly improves the understanding of the almost - prime distribution of high - degree polynomials evaluated at prime numbers by improving sieve method techniques.