Precise tail behaviour of some Dirichlet series (Q6592153)

The authors consider a sequence \(\eta_k, k\ge 1\) of independent copies of a random variable \(\eta\) with zero mean and finite variance. Additionally, they assume that \(\eta\) is bounded above, \(\eta\le b\) a.s. for some \(b>0\). Considering different types of the asymptotic behaviour of the probability \(\mathbb{P}\{\eta\in[b-x, b]\}\) as \(x \to 0+\), they derive precise tail asymptotics of the random Dirichlet series \(S(\alpha)=\sum_{k\ge1} k^{-\alpha}\eta_k\) for \(\alpha\in (1/2, 1]\). In principle, the asymptotic behaviour of the distribution tail of \(S(\alpha)\) can be deduced from the asymptotic behaviour of the corresponding density. However, it is mentioned that the authors' approach used in proving the precise tail asymptotics of \(S(\alpha)\) has its own merits and may be used for investigating precise distributional tail behaviour of other infinite weighted sums of independent identically distributed random variables. One of its advantages is that absolute continuity of the distribution of a sum is not a priori required.