I am trying to find an algorithm for this. You can imagine each set $(S_1, S_2, \ldots, S_n)$ as points with different colour. Also it isn't necessarily $|S_1|=|S_2|=\cdots=|S_n|$.
For $n=1$ we trivially have a single (random) point.
For $n=2$ we have two sets of points $S, Q$ and we seek to find the (distance of the) closest pair of points $p, q$ such that $p\in S$ and $q \in Q$. I have also found an efficient algorithm for this, using voronoi diagrams.
For $n>2$ things get tricky. I have no clue where should I head to. Let's say we have $x$ red, $y$ green and $z$ blue points laid down in an Euclidian plane. How do we find the minimum distance of a route passing for one red, one green and one blue point?
Is this some special case of the Traveling Purchaser Problem?