It is worthwhile that interest will focus on lower and upper records, simultaneously. In this paper, we intend to investigate some properties of such a records in two different plans. To this end, let $X_{i,j}$ $ (1\\\\leq i\\\\leq k,j \\\\geq 1)$ be independent sequences of random variables and from each, $m$ upper and $n$ lower record values are extracted, jointly. We consider the problems of constructing the distribution-free confidence intervals based on simultaneous record data. We compare the results in a single sequence ($k=1$) and in a multi-sequence ($k > 1$) of observations. It is shown that how one can choose the minimum number of required records.

Keywords