SUMMARYWhen the individual PDFs of closely-spaced random variables such as natural frequencies of a structure overlap, generation of sample sets by assuming the frequencies to be independent random variables can lead to incorrect sets of frequencies in the sense that the frequencies do not remain as ordered sets. Rejection of such disordered sample sets results in individual density functions that are significantly different from the distributions initially assumed for sampling each random variable. One way to overcome this constraint in the simulation of an ordered set of random variables is to consider them in an implicit manner using a joint PDF. In this paper, we present a formulation for a joint density function that is developed using fundamental probability approaches. The formulation ensures that the sampled random variables always remain as ordered sets and maintain the individual density functions for each variable. Application of the proposed formulation is illustrated for cases with not just two closely-spaced variables but also for a case with multiple closely-spaced variables such that the PDFs of more than two random variables overlap with each other. An expression is presented to determine the exact number of terms needed in the formulation. However, it is also illustrated that only two terms are sufficient in most applications even when the exact number of terms needed is very high.