Yes it is possible due to finite integer size wraparound.

It is a frequently asked question. Please check the FAQ portion of this forum.

I have checked the FAQ,But there is no explication for the reason that you have authorized to have a negative value of cardinality.
I don't imagine some one who look-up for a negative cardinality of a SET, and even in Logic there is no negative cardinality , it was that my question .
thanks for the request,you're always here for need ...

You claim you found Alloy seems to mistakenly allow a set to have negative cardinality. That's not it.

#Set is not the cardinality of the set

#Set is the integer truncation result of the cardinality of the set

While the cardinality of a set in Alloy is never negative, the integer truncation of that number certainly can be.
For example, if you chose to run a command or check an assertion where the integer
ranges from -8 to +7 (if you set integer bit with of 4), then if a set has 8 elements, then #set evaluates to -8.
(-8, -7, -6, .., -2, -1, 0, 1, 2, .., 6, 7, -8, -7, -6, .., -2, -1, 0, 1, 2, ... so on and so on)

The relevant answers are http://alloy.mit.edu/community/node/204
and http://alloy.mit.edu/community/node/23
in particular, talk about since Alloy is a declarative language, there's no way
to throw a "runtime exception"; every expression must have some legal meaning.
So when we add 2 numbers in Alloy, we must define a precise mathematical meaning
to the result, even if the 2 numbers happen to both be the maximum positive integers, for example.
The meaning we chose is the standard 2s complement arithmetic.

Thanks ...