The components of maximum (Q810993)

Denote \(X=\max \{X_ 1,X_ 2\}\) where \(X_ 1\) and \(X_ 2\) are independent random variables. The distribution functions \(F_ 1\) of \(X_ 1\) and \(F_ 2\) of \(X_ 2\) are called the max-component of the distribution function F of X. The main result of the paper is the following Theorem. For the distribution function \(F_ 1\) to be a max-component of the distribution function F it is necessary and sufficient that \[ F_ 1(x+\ell)F(x)\leq F(x+\ell)F_ 1(x),\quad x\in R,\quad \ell \geq 0, \] and, if lext F\(=-\infty\), that \[ \lim_{x\to -\infty}F(x)/F_ 1(x)=0. \] There are given criteria when an absolutely continuous (purely discrete, purely singular) distribution function has the max-components with the same property. - There are a few misprints in the paper.