No label defined (Q4234474)

For Part II, see Lith. Math. J. 37, No. 2, 162-170 (1997); translation from Liet. Mat. Rink. 37, No. 2, 212-223 (1997; Zbl 0892.11030), for Part I, see Zbl 0977.11042 above. Let \(G(n)=g_1(a_1 n+b_1)\cdots g_s(a_s n+b_s)\), where the \(g\)'s are complex-valued multiplicative functions, the \(a\)'s are fixed positive integers and the \(b\)'s are fixed integers. Set \(M_x(G)=x^{-1}\sum_{n\leq x} G(n)\). The object of the paper is the proof of an upper bound for \(M_x(G)-P(x)\) under suitable assumptions on the functions \(g\), \(P\) being the expected main term, which is given explicitly in terms of the above parameters. Among the consequences, we quote the proof that the distribution functions associated to additive functions belonging to a wide class converge weakly to some limiting distribution whose characteristic function can be explicitly computed. The fourth part of the series [Publ. Math. (Debrecen) 52, 659-681 (1998; Zbl 0912.11036)] deals with similar problems in which the averages are computed over primes. For the entire collection see [Zbl 0896.00021].