Man, Woman, Dog, seeking stable relationship.

689 Views Asked by At

There is a classic problem in combinatorics dealing with a stable pairing between a set of men and a set of women as spouses. (Gale-Shapely algorithm) http://en.wikipedia.org/wiki/Stable_marriage_problem

Is there are equivalent algorithm for determining the stability of a pairing where women choose men, men choose dogs, and dogs choose women? Are we guaranteed to arrive at a stable matching consisting of a man, a woman, and a dog?

I will define stable to mean that if we have two triples $(m_1, w_1, d_1$) and ($m_2, w_2, d_2$), then we cannot have that $w_1$ desires to be with $m_2$ more than with $m_1$ and $m_2$ desires to own $d_1$ more than he desires to own $d_2$. We ignore the preference of the dogs in this case. So instability is defined to mean that a member of one triple, call it $T_1$, desires another partnering, and the person or dog with whom they wish to join also desires to leave their own pairing to join with the third member of $T_1$.

It would seem like finding a stable matching may be much more challenging than the original man woman problem. Are there any proofs of impossibility that we know of?

1

There are 1 best solutions below

0
On

While the question itself is not as hard and fast as Gale-Shapely, there are methods for finding a "stable-enough" matching. Similar to many approximate solutions to the traveling salesman problem, many times all we need is a good approximation.

The 1st and 3rd references above give that finding a stable matching is an NP-complete problem. As I discover more on the subject I will continue to update this answer perhaps.

But, for now, the simple answer is that finding a stable matching is an NP-complete problem.