[seqfan] Most efficient algorithm for selecting divisors of an abundant number?

[Alonso Del Arte]

What would be the most efficient algorithm for selecting divisors of an
abundant number that add up to that number such that the smallest divisors
are used and the fewest divisors are skipped over for larger divisors?

In the case of the Erdős-Nicolas numbers (A194472) the answer is easy: just
select the first k divisors.

In the case of the weird numbers (A006037) the answer is just as easy: no
selection is possible.

But with other abundant numbers, how do you decide which divisors to skip
over, other than trying each possible selection and weighing it against the
goals?

For example, with 30, I'm not convinced 1 + 3 + 5 + 6 + 15 (skipping over 2
and 10) is the best selection.

Al