The Hardy-Ramanujan and Halász inequalities for shifted primes (Q1909790)

The Hardy-Ramanujan inequality referred to in the title gives a uniform upper bound for the quantity \(N(m, x,g)= \#\{ n\leq x: g(n)= m\}\), where the function \(g(n)\) is either \(\omega (n)\), the number of distinct prime factors of \(n\), or \(\Omega (n)\), the number of prime factors of \(n\) counted with multiplicity: in the first case, the bound \[ N(m, x, g)\leq {{c_1 x(\log \log x+ c_2)^{m-1}} \over {(m- 1)! \log x}} \tag{1} \] holds uniformly for \(x\geq 2\) and all positive integers \(m\) with suitable constants \(c_1\) and \(c_2\); in the second case, (1) holds in the range \(m\leq (2- \delta) \log \log x\) for any given positive constant \(\delta> 0\). \textit{G. Halász} [Acta Math. Acad. Sci. Hung. 23, 425-432 (1972; Zbl 0255.10046)] generalized the Hardy-Ramanujan upper bound to functions of the type \(g(n)= \Omega(n, {\mathcal P})\), where \({\mathcal P}\) is an arbitrary set of primes, and \(\Omega (n, {\mathcal P})\) denotes the number of prime factors of \(n\), counted with multiplicity and belonging to the set \({\mathcal P}\); he also obtained comparable lower bound estimates for \(N(m, x, g)\). In the paper under review, the author proves analogs of these inequalities for the counting functions \(P(m, x, g)= \#\{p \leq x: p\) prime, \(g(p+a)= m\}\), where \(a\) is a fixed, non-zero integer. The author's bounds for \(P(m, x, g)\) are of the same type as those in the Hardy-Ramanujan and Halász inequalities, except that they involve (as expected) an additional factor \(\log x\) in the denominator. For example, in the case \(g(n)= \omega (n)\), he shows that the inequalities \[ P(m, x, g)\leq {{c_3 x(\log \log x+ c_4 )^{m-1}} \over {(m-1)! \log^2 x}} \quad \text{and} \quad P(m, x, g)+ P(m+ 1, x, g)\geq {{c_5 x(\log \log x)^{m-1}} \over {(m- 1)! \log^2 x}} \] hold uniformly for \(x\geq |a|\) and \(m\leq b\log \log x\), for any given \(b> 1\), with constants \(c_i\) depending at most on the parameters \(a\) and \(b\). The proofs rely largely on methods and results from sieve theory. [Also submitted to MR].