> Instance: Collection C of subsets of a finite set S, positive integer K <= |S| * |C|, |S| = O(n) and |C| = O(n). For each subset c_i in C, we can choose some items in c_i to delete.
> Question: Is there a way to remove at most K items from C such that after removal all resulting sets in C are distinct. For example, let S={1,2,...n}, C={c_1={1,2,4}, c_2={1,2,4}, c_3={2,4}} and K = 1. We can remove {2} from c_1 and the resulting sets are all distinct.
> 
> 这个问题有什么高效的算法没有？ 抑或有人能给出一个np-complete的证明么？