Remarks on domination of maxima (Q1871318): Difference between revisions

The author considers the sequence \(\{X_n\}_{n\geq 1}\) of independent random variables having the common continuous distribution function \(F\), and two sequences \((a_n)_{n\geq 1}\), \((b_n)_{n\geq 1}\) of real constants. Let also \(M_n= \max\{X_1,X_2,\dots, X_n\}\), \(n\in\mathbb{N}\). Then the aim of the paper is to investigate proper conditions on \(F\), \(a_n\), \(b_n\), under which one can establish the asymptotic behaviour of the probability \({\mathbf P}(M_n\leq a_n X_1+ b_n)\). Connections to similar results in the literature are pointed out. E.g., when \(F\) is tail equivalent to the standard Gaussian distribution function on \(\mathbb{R}\), and both sequences \(a_n> 0\) and \(b_n\in\mathbb{R}\) depend on \(n\), then the asymptotic evaluation of the probability \({\mathbf P}(M_n\leq a_n X_1+ b_n)\) is derived as a generalization of a result of \textit{Y. Rinott} and \textit{V. Rotar} [Decis. Econ. Finance 23, No. 1, 15-29 (2000; Zbl 0985.60024)].