Prime polynomials in short intervals and in arithmetic progressions
Efrat Bank,Lior Bary-Soroker,Lior Rosenzweig
DOI: https://doi.org/10.1215/00127094-2856728
2015-02-01
Duke Mathematical Journal
Abstract:In this paper we establish function field versions of two classical conjectures on prime numbers. The first says that the number of primes in intervals (x,x+xϵ] is about xϵ/logx. The second says that the number of primes p<x in the arithmetic progression p≡a(modd), for d<x1−δ, is about π(x)ϕ(d), where ϕ is the Euler totient function. More precisely, for short intervals we prove: Let k be a fixed integer. Then πq(I(f,ϵ))∼#I(f,ϵ)k,q→∞ holds uniformly for all prime powers q, degree k monic polynomials f∈Fq[t] and ϵ0(f,q)≤ϵ, where ϵ0 is either 1k, or 2k if p∣k(k−1), or 3k if further p=2 and degf'≤1. Here I(f,ϵ)={g∈Fq[t]∣deg(f−g)≤ϵdegf}, and πq(I(f,ϵ)) denotes the number of prime polynomials in I(f,ϵ). We show that this estimation fails in the neglected cases. For arithmetic progressions we prove: let k be a fixed integer. Then πq(k;D,f)∼πq(k)ϕ(D),q→∞, holds uniformly for all relatively prime polynomials D,f∈Fq[t] satisfying ‖D‖≤qk(1−δ0), where δ0 is either 3k or 4k if p=2 and (f/D)' is a constant. Here πq(k) is the number of degree k prime polynomials and πq(k;D,f) is the number of such polynomials in the arithmetic progression P≡f(modd). We also generalize these results to arbitrary factorization types.
mathematics