2. Lecture 23: Stable Marriage ( Based on Lectures of Steven Rudich of CMU and Amit Sahai of Princeton) Shang-Hua Teng

3. Dating Scenario There are n “men” and n “women” Each woman has her own ranked preference list of ALL the men Each man has his own ranked preference list of ALL the women The lists have no ties

4. 3,5,2,1,4 1 5,2,1,4,3 4,3,5,1,2 3 1,2,3,4,5 4 2,3,4,1,5 5 1 3,2,5,1,4 2 1,2,5,3,4 3 4,3,2,1,5 4 1,3,4,2,5 5 1,2,4,5,3 2

5. The Matching (Marriage) Problem Question: How do we pair them off? Which criteria come to mind?

6. There is more than one notion of what constitutes a “good” pairing. Maximizing the number of people who get their first choice Maximizing average satisfaction Maximizing the minimum satisfaction

7. Couple of “Mutually Cheating Hearts” Suppose we pair off all the men and women. Now suppose that some man and some woman prefer each other to the people they married. They will be called a couple of MCHs.

8. Why be with them when we can be with each other?

9. Stable Pairings A pairing of men and women is called stable if it contains no couple of MCHs.

10. Stability is Primary. Any list of criteria for a good pairing must include stability. (A pairing is doomed if it contains a shaky couple.) Any reasonable list of criteria must contain the stability criterion.

11. The study of stability will be the subject of the entire lecture. We will: –Analyze an algorithm that looks a lot like dating in the early 1950’s naked mathematical truth –Discover the naked mathematical truth about which sex has the romantic edge –Learn how the world’s largest, most successful dating service operates

12. How do we find a stable pairing?

13. Wait! There is a more primary question!

14. How do we find a stable pairing? How do we know that a stable pairing always exists?

15. A “Traditional” Marriage Algorithm Choose the best among those who propose, “get a commitment” Propose to the best who is still available

16. A Today’s Dating Algorithm

17. Traditional Marriage Algorithm For each day each single man does the following: –Morning Each bachelorette sits in her favorite coffee house Each single men proposes to the best bachelorette to whom he has not yet proposed –Afternoon (for those bachelorettes with at least one suitor) “Maybe, come back tomorrow, and ask me again”To today’s best suitor: “Maybe, come back tomorrow, and ask me again” “Not me, but good luck”To any others: “Not me, but good luck” –Evening Each single man is ready to propose next bachelorette on his list The day WILL COME that no bachelor is rejected Each bachelorette marries the last bachelor to whom she said “maybe”

18. Does the Traditional Marriage Algorithm always produce a stable pairing?

19. Wait! There is a more primary question!

20. Does the TMA work at all? –Does it eventually stop? –Is everyone married at the end?

21. Theorem: The TMA always terminates in at most n 2 days Each day that at least one Bachelor gets a “No”. There can be at most n 2 “No”s.

22. Does the TMA work at all? –Does it eventually stop? –Is everyone married at the end?

23. Improvement Lemma: If a bachelorette has a committed suitor, then she will always have someone at least as good the next day (and on the final day). –She would only let go of him in order to “maybe” someone better –She would only let go of that guy for someone even better –AND SO ON.............

24. Corollary: Each bachelorette will marry her absolute favorite of the bachelors who proposed to her during the TMA

25. Lemma: No bachelor can be rejected by all the bachelorettes Proof by contradiction. Suppose Bob is rejected by all the women. At that point: Each women must have a suitor other than Bob (By Improvement Lemma, once a woman has a suitor she will always have at least one) The n women have n suitors, Bob not among them. Thus, there must be at least n+1 men! Contradicti on

26. Great! We know that TMA will terminate and produce a pairing. But is it stable?

27. Theorem: The pairing produced by TMA is stable. –This means Bob likes Mia more than his partner, Alice. –Thus, Bob proposed to Mia before he proposed to Alice. –Mia must have rejected Bob for someone she preferred. Luke –By the Improvement lemma, she must like her parnter Luke more than Bob. Bob Alice Mia Luke –Proof by contradiction: Suppose Bob and Mia are a couple of MCHs. Contradiction!

28. Opinion Poll Who is better off in traditional dating, the men or the women?

29. Forget TMA for a moment How should we define what we mean when we say “the optimal woman for Bob”? Flawed Attempt: “The woman at the top of Bob’s list”

30. The Optimal Match A man’s optimal match is the highest ranked woman for whom there is some stable pairing in which they are matched She is the best woman he can conceivably be matched in a stable world. Presumably, she might be better than the woman he gets matched to in the stable pairing output by TMA.

31. The Pessimal Match A man’s pessimal match is the lowest ranked woman for whom there is some stable pairing in which the man is matched to. She is the least ranked woman he can conceivably get to be matched to in a stable world.

32. Dating Dilemmas A pairing is man-optimal if every man gets his optimal match. This is the best of all possible stable worlds for every man simultaneously. A pairing is man-pessimal if every man gets his pessimal match. This is the worst of all possible stable worlds for every man simultaneously.

33. Dating Dilemmas A pairing is woman-optimal if every woman gets her optimal match. This is the best of all possible stable worlds for every woman simultaneously. A pairing is woman-pessimal if every woman gets her pessimal match. This is the worst of all possible stable worlds for every woman simultaneously.

34. The Naked Mathematical Truth! The Traditional Marriage Algorithm always produces a man-optimal, and woman-pessimal pairing.

35. Theorem: TMA produces a man-optimal pairing –Suppose not: i.e. that some man gets rejected by his optimal match during TMA. –In particular, let’s say Bob is the first man to be rejected by his optimal match Mia: Let’s say she said “maybe” to Luke, whom she prefers. –Since Bob was the only man to be rejected by his optimal match so far, Luke must like Mia at least as much as his optimal match.

36. We are assuming that Mia is Bob’s optimal match Mia likes Luke more than Bob. Luke likes Mia at least as much as his optimal match. We now show that any pairing S in which Bob marries Mia cannot be stable (for a contradiction). Suppose S is stable: –Luke likes Mia more than his partner in S Luke likes Mia at least as much as his best match, but he is not matched to Mia in S –Mia likes Luke more than her partner Bob in S Luke Mia Contradiction!

37. We’ve shown that any pairing in which Bob marries Mia cannot be stable. –Thus, Mia cannot be Bob’s optimal match (since he can never marry her in a stable world). –So Bob never gets rejected by his optimal match in the TMA, and thus the TMA is man-optimal. We are assuming that Mia is Bob’s optimal match. Mia likes Luke more than Bob. Luke likes Mia at least as much as his optimal match.

38. Theorem:The TMA pairing is woman-pessimal. We know it is man-optimal. Suppose there is a stable pairing S where some woman Alice does worse than in the TMA. Let Luke be her partner in the TMA pairing. Let Bob be her partner in S. –By assumption, Alice likes Luke better than her partner Bob in S –Luke likes Alice better than his partner in S We already know that Alice is his optimal match ! Luke Alice Contradiction!

39. Lessons???

40. The largest dating service in the world uses a computer to run TMA !

41. “The Match”: Doctors and Medical Residencies –Each medical school graduate submits a ranked list of hospitals where he/she wants to do a residency –Each hospital submits a ranked list of newly minted doctors that it wants –A computer runs TMA (extended to handle polygamy) –Until recently, it was hospital-optimal

42. An Instructive Variant: Team Projects 1 3 2,3,4 1,2,4 2 4 3,1,4 *,*,*

43. An Instructive Variant: Team Project 1 3 2,3,4 1,2,4 2 4 3,1,4 *,*,*

44. An Instructive Variant: Team Project 1 3 2,3,4 1,2,4 2 4 3,1,4 *,*,*

45. An Instructive Variant: Team Project 1 3 2,3,4 1,2,4 2 4 3,1,4 *,*,*

46. No Stable Teaming Possible 1 3 2,3,4 1,2,4 2 4 3,1,4 *,*,*

47. Stability Amazingly, stable pairings do always exist in the bipartite relations. Insight: We must make sure our arguments do not apply to the class project case, since we know all arguments must fail in that case.

48. REFERENCES D. Gale and L. S. Shapley, College admissions and the stability of marriage, American Mathematical Monthly 69 (1962), 9-15 Dan Gusfield and Robert W. Irving, The Stable Marriage Problem: Structures and Algorithms, MIT Press, 1989

49. History: When Markets fail 1900 –Idea of hospitals having residents (then called “interns”) Over the next few decades –Intense competition among hospitals for an inadequate supply of residents Each hospital makes offers independently Process degenerates into a race. Hospitals steadily advancing date at which they finalize binding contracts

50. History: When Markets fail 1944 Absurd Situation. Appointments being made 2 years ahead of time! –All parties were unhappy –Medical schools stop releasing any information about students before some reasonable date –Did this fix the situation?

51. History: When Markets fail 1944 Absurd Situation. Appointments being made 2 years ahead of time! –All parties were unhappy –Medical schools stop releasing any information about students before some reasonable date –Offers were made at a more reasonable date, but new problems developed

52. History: When Markets fail 1945-1949 Just As Competitive –Hospitals started putting time limits on offers –Time limit gets down to 12 hours –Lots of unhappy people –Many instabilities resulting from lack of cooperation

53. History: When Markets fail 1950 Centralized System –Each hospital ranks residents –Each resident ranks hospitals –National Resident Matching Program produces a pairing Whoops! The pairings were not always stable. By 1952 the algorithm was the TMA (hospital-optimal) and therefore stable.

54. History Repeats Itself! NY TIMES, March 17, 1989 –“The once decorous process by which federal judges select their law clerks has degenerated into a free-for-all in which some of the judges scramble for the top law school students...” –“The judges have steadily pushed up the hiring process...” –“Offered some jobs as early as February of the second year of law school...” –“On the basis of fewer grades and flimsier evidence...”

55. NY TIMES –“Law of the jungle reigns.. “ –The association of American Law Schools agreed not to hire before September of the third year... –Some extend offers from only a few hours, a practice known in the clerkship vernacular as a “short fuse” or a “hold up”. –Judge Winter offered a Yale student a clerkship at 11:35 and gave her until noon to accept... At 11:55.. he withdrew his offer