Intermediate- and extreme-sum processes (Q1185784)

Let \(X_ 1,X_ 2,\dots\) be i.i.d. random variables, for each integer \(n\geq 1\) let \(X_{1,n}\leq\dots\leq X_{n,n}\) denote the order statistics of \(X_ 1,\dots,X_ n\), and let \(k_ n\) be a sequence of positive numbers such that \(k_ n\to\infty\) and \(k_ n/n\to 0\) as \(n\to\infty\). For \(0<a<b\) consider the sum process \[ I_ n(a,t)=\sum_{i=\lceil ak_ n\rceil+1}^{\lceil tk_ n\rceil} X_{n+1-i,n}, \qquad a\leq t\leq b, \] of intermediate order statistics and the sum process \[ E_ n(t)=\begin{cases} \sum_{i=1}^{\lceil tk_ n\rceil} X_{n+1-i,n}, & 1/k_ n\leq t\leq b,\\ 0, & 0\leq t<1/k_ n, \end{cases} \] of extreme order statistics. The weak convergence behavior of the processes \(I_ n\) and \(E_ n\), with suitable centering and norming and possibly along a subsequence of the sequence \(\{n\}\) of positive integers, is studied.