On primes in arithmetic progressions (Q5949480)

Let \(P(q,a)\) denote the least prime in the arithmetic progression \(p\equiv a\pmod q\), where \((q,a)=1\). Linnik proved that \(P(q,a)\ll q^L\) for an absolute constant \(L\), and \textit{D. R. Heath-Brown} [Proc. Lond. Math. Soc. (3) 64, 265-338 (1992; Zbl 0739.11033)] showed that \(L=5{\cdot}5\) is an admissible value. The author proves the following theorem: Let \(K>32/17\) and \(A,B>0\). Then, for large \(Q\) and \(0<|a|<\log^BQ\), we have \(P(q,a)\ll q^K\) with at most \(O(Q\log^{-A}Q)\) exceptional values \(q\) in \(Q<q\leq 2Q\) with \((q,a)=1\). The exponent \(32/17\) improves \(25/13\), which was due to \textit{R. C. Baker} and \textit{G. Harman} [Analytic Number Theory, Vol. 1, Prog. Math. 138, 39-103 (1996; Zbl 0853.11078)]. Besides the work of \textit{E. Bombieri, J. B. Friedlander} and \textit{H. Iwaniec} [J. Am. Math. Soc. 2, 215-223 (1989; Zbl 0674.10036)], the author also applies a sieve identity associated with the incomplete sum \(\sum_{d|n,d<\sqrt n}\mu(d)\), which was first considered by Vinogradov and seems to be particularly relevant and effective in this problem.