Multiplicative functions in arithmetic progressions (Q743704)

As it is well-known, for any \(\varepsilon > 0\) there exists an \(A\) such that \[ \left| \pi(x;q,a)-\frac{\pi(x)}{\phi(q)}\right| \leq \varepsilon\frac{\pi(x)}{\phi(q)} \] for all \((a,q)=1\) and \(q\leq x^{1/A}\), except possibly for those \(q\) which are multiples of some so-called exceptional modulus \(r\) (depending only on \(\varepsilon\) and \(x\)). (Here \(\phi\) is Euler's function.) It is widely believed that in fact no exceptional moduli exist. Previously, exceptional moduli have been investigated in terms of \(L\)-functions (see results of \textit{H. Daboussi} and \textit{H. Delange} [C. R. Acad. Sci., Paris, Sér. A 278, 657--660 (1974; Zbl 0292.10034)], and previous work by the last two authors). The main motivation of the paper is to develop a similar theory for multiplicative functions \(f(n)\) (with values inside or on the unit circle) in arithmetic progressions, however, without using \(L\)-functions. The authors provide estimates for the mean values \[ \frac{1}{x/q} \underset{n\equiv a\pmod{q}}{\sum\limits_{n\leq x}} f(n) \] for such functions \(f(n)\), where \((a,q)=1\). Some results in the paper overlap with certain results of \textit{P. D. T. A. Elliott} [J. Lond. Math. Soc., II. Ser. 66, No. 1, 14--28 (2002; Zbl 1020.11059); in: Number theory for the millennium I. Proceedings of the millennial conference on number theory, Urbana-Champaign, USA, 2000. Natick, MA: A K Peters. 393--418 (2002; Zbl 1032.11040)]. The connections are clarified, and it is argued that the proofs in the present paper are different and easier.