Optimisation under uncertainty: using multiple samples from same distribution

[sitadrost]

For a design optimisation problem with uncertain variable values (e.g. dimensions with tolerance, material properties with uncertainty), Dakota's Optimization Under Uncertainty capabilities seem ideally suited, so I'm trying to learn about the available methods and how to use them. I was wondering:

Say my problem has one uncertain design variable that occurs twice (e.g. a 3-bar truss with two (nominally) identical bars, such as in this tutorial), how can I tell Dakota to consider both instances of this variable as distinct/independent samples from the same distribution? I assume that this doesn't happen automatically?