If I have a set of words that form a text, say text $a$, and a set of texts, I then calculate the similarity between text $a$ and each text in the set. Then, I get a multiset.
For example:
$$s = \{2, 3, 2, 4, 6, 4, 3, 6\}$$
How can I uniformly distribute it?
What I actually want is to randomly select from this multiset, but I don't know how to deal with the multiset and to form it.