Uniform bounds on the relative error in the approximation of upper quantiles for sums of arbitrary independent random variables (Q501825)

Let \(\tilde{X}_1,\dots,\tilde{X}_n\) be independent random variables having arbitrary given distributions. The authors' aim in this paper is to study the upper quantile function defined by \(\tilde{s}\equiv\sup[s:\operatorname{P}(\tilde{S}_n\geq s)\geq\frac1y]\), where \(y>1\) and \(\tilde{S}_n=\sum^n_{j=1}\tilde{X}_j\). They construct a computable quantity \(Q_y\) based on the marginal distributions \(x_1,\dots, x_n\) to produce upper and lower bounds for \(S_y\) and \(s^\ast_y\) -- the upper \(\frac 1y\)th of \(S_n=\sum^n_{j=1}X_j\) and \(S^\ast=\sup_{1\leq k\leq n}S_n\), respectively.