Content deleted Content added
Arthur Frayn (talk | contribs) ('''SMP''') |
|||
Line 40:
''Proof that the Gale-Shapley pairing is male-optimal:'' We go by contradiction. Suppose the pairing generated by the Gale-Shapley algorithm were not male-optimal. Let T be the first iteration where a man is rejected by his optimal partner. Let man M be one man who is rejected by his optimal partner, woman W, during this iteration. Then woman W must have rejected man M for some other man, man Z. Further, since this is the first iteration where a man is rejected by his optimal partner, woman W must be at least as high on man Z's preference list as his optimal partner. Also note that, since woman W is man M's optimal partner, there exists some stable pairing S where man M is paired with woman W. Then, in pairing S, woman W prefers man Z to man M (since she rejected M for Z in the Gale-Shapley algorithm) and man Z prefers woman W to his partner in S, since he likes woman W at least as much as his optimal partner and, by our definition of optimal, he could not be paired with anyone better than this. But then, by definition, this pairing is unstable, since man Z and woman W prefer each other to their partners in S. We have reached a contradiction, so the Gale-Shapley algorithm must produce a male-optimal pairing. <math>\Box</math>
''Proof that the Gale-Shapley pairing is female-pessimal:'' We can show this using the fact that
==Applications==
|