2
$\begingroup$

Given integers $a_1,\ldots,a_k>0$ and $b_1,\ldots,b_k$, consider the polynomial $f(x) = \prod_{i=1}^{k} (a_i x +b_i) \in \mathbb{Z}[x]$. Suppose that $\{ a_i x+b_i\}_{i=1}^{k}$ are pairwise coprime polynomials of degree 1.

It is known (see e.g. Chapter 10 of `Sieve Methods' by Halberstam and Richert) that if $\gcd(f(1),f(2),f(3),\ldots)=1$, then for any sufficiently large $r$ (in terms of $k$), there are infinitely many $n$-s such that $f(n)$ is a product of at most $r$ primes. In fact, there are at least $\Omega_{a_i,b_i,k,r}(\frac{x}{\log^k x})$ such $n$-s up to $x$, and one may take $r > C k \log (k+1)$ for some absolute constant $C$.

Is it possible to prove a similar result with modular constraints on the prime factors of $F(n)$? E.g., is it known that given $k$, there is an integer $r$ with the following property: $f(x)=\prod_{i=1}^{k} (4 a_i x + 1)$ is a product of at most $r$ primes of the form $1 \bmod 4$ for infinitely many $n$?

Note that the condition that the prime factors of $f(n)$ are $1 \bmod 4$ is consistent with the congruence $f(n) \equiv 1 \bmod 4$.

$\endgroup$
2
  • 1
    $\begingroup$ If a prime $p$ divides $f(n)=n^2 +1$, then $-1$ is a square mod $p$, and thus $p$ cannot be congruent to $3$ mod $4$. Could you clarify whether you're interested in cases where reciprocity laws do not give such additional congruence constraints for free? $\endgroup$
    – GNiklasch
    Commented Oct 17, 2018 at 15:32
  • $\begingroup$ @GNiklasch: That's a great observation, I have edited the question accordingly. $\endgroup$ Commented Oct 17, 2018 at 15:42

0

You must log in to answer this question.