hri: All stable matchings in the hospital/residents problem with...

In matchingMarkets: Analysis of Stable Matchings

All stable matchings in the hospital/residents problem with incomplete lists

Description

Finds all stable matchings in either the hospital/residents problem (a.k.a. college admissions problem) or the related stable marriage problem. Dependent on the problem, the results comprise the student and college-optimal or the men and women-optimal matchings. The implementation allows for incomplete preference lists (some agents find certain agents unacceptable) and unbalanced instances (unequal number of agents on both sides). The function uses the Prosser (2014) constraint encoding based on either given or randomly generated preferences.

Usage

hri(
  nStudents = ncol(s.prefs),
  nColleges = ncol(c.prefs),
  nSlots = rep(1, nColleges),
  s.prefs = NULL,
  c.prefs = NULL,
  s.range = NULL,
  c.range = NULL,
  randomization = NULL,
  seed = NULL,
  check_consistency = TRUE,
  verbose = FALSE,
  ...
)

Arguments

nStudents	integer indicating the number of students (in the college admissions problem) or men (in the stable marriage problem) in the market. Defaults to ncol(s.prefs).
nColleges	integer indicating the number of colleges (in the college admissions problem) or women (in the stable marriage problem) in the market. Defaults to ncol(c.prefs).
nSlots	vector of length nColleges indicating the number of places (i.e. quota) of each college. Defaults to rep(1,nColleges) for the marriage problem.
s.prefs	matrix of dimension nColleges x nStudents with the jth column containing student j's ranking over colleges in decreasing order of preference (i.e. most preferred first).
c.prefs	matrix of dimension nStudents x nColleges with the ith column containing college i's ranking over students in decreasing order of preference (i.e. most preferred first).
s.range	range of two intergers s.range = c(s.min, s.max), where s.min < s.max. Produces incomplete preference lists with the length of each student's list randomly sampled from the range [s.min, s.max]. Note: interval is only correct if either c.range or s.range is used.
c.range	range of two intergers c.range = c(c.min, c.max), where c.min < c.max. Produces incomplete preference lists with the length of each college's list randomly sampled from the range [c.min, c.max]. Note: interval is only correct if either c.range or s.range is used.
randomization	determines at which level random lottery numbers for student priorities are drawn. The default is randomization = "multiple", where a student's priority is determined by a separate lottery at each college (i.e. local tie-breaking). For the second variant, randomization = "single", a single lottery number determines a student's priority at all colleges (i.e. global tie breaking).
seed	integer setting the state for random number generation.
check_consistency	Performs consicentcy checks (Checks if there are columns in the preference matrices that only contains zeros and drops them and checks the matrixes for consistencies if they are given by characters). Defaults to TRUE but changing it to FALSE might reduce the running-time for large problems.
verbose	logical. When set to TRUE, writes information messages on the console (recommended). Defaults to FALSE, which suppresses such messages.
...	.

Value

hri returns a list of the following elements.

s.prefs.smi	student-side preference matrix for the stable marriage problem with incomplete lists (SMI).
c.prefs.smi	college-side preference matrix for the stable marriage problem with incomplete lists (SMI).
s.prefs.hri	student-side preference matrix for the college admissions problem (a.k.a. hospital/residents problem) with incomplete lists (HRI).
c.prefs.hri	college-side preference matrix for the college admissions problem (a.k.a. hospital/residents problem) with incomplete lists (HRI).
matchings	edgelist of matched students and colleges, inculding the number of the match (matching) and two variables that indicate the student-optimal match (sOptimal) and college-optimal match (cOptimal)

.

Minimum required arguments

hri requires the following combination of arguments, subject to the matching problem.

nStudents, nColleges

Marriage problem with random preferences.

s.prefs, c.prefs

Marriage problem with given preferences.

nStudents, nSlots

College admissions problem with random preferences.

s.prefs, c.prefs, nSlots

College admissions problem with given preferences.

Author(s)

Thilo Klein

References

Gale, D. and L.S. Shapley (1962). College admissions and the stability of marriage. The American Mathematical Monthly, 69(1):9–15.

Morizumi, Y., T. Hayashi and Y. Ishida (2011). A network visualization of stable matching in the stable marriage problem. Artificial Life Robotics, 16:40–43.

Prosser, P. (2014). Stable Roommates and Constraint Programming. Lecture Notes in Computer Science, CPAIOR 2014 Edition. Springer International Publishing, 8451: 15–28.

Examples

## -----------------------
## --- Marriage problem 

## 7 men, 6 women, random preferences:
 hri(nStudents=7, nColleges=6, seed=4)

## 3 men, 2 women, given preferences:
 s.prefs <- matrix(c(1,2, 1,2, 1,2), 2,3)
 c.prefs <- matrix(c(1,2,3, 1,2,3), 3,2)
 hri(s.prefs=s.prefs, c.prefs=c.prefs)

## 3 men, 2 women, given preferences:
 s.prefs <- matrix(c("x","y", "x","y", "x","y"), 2,3)
 colnames(s.prefs) <- c("A","B","C")
 c.prefs <- matrix(c("A","B","C", "A","B","C"), 3,2)
 colnames(c.prefs) <- c("x","y")
 hri(s.prefs=s.prefs, c.prefs=c.prefs)

## --------------------------------
## --- College admission problem 

## 7 students, 2 colleges with 3 slots each, random preferences:
 hri(nStudents=7, nSlots=c(3,3), seed=21)

## 7 students, 2 colleges with 3 slots each, given preferences:
 s.prefs <- matrix(c(1,2, 1,2, 1,NA, 1,2, 1,2, 1,2, 1,2), 2,7)
 c.prefs <- matrix(c(1,2,3,4,5,6,7, 1,2,3,4,5,NA,NA), 7,2)
 hri(s.prefs=s.prefs, c.prefs=c.prefs, nSlots=c(3,3))
 
## 7 students, 2 colleges with 3 slots each, given preferences:
 s.prefs <- matrix(c("x","y", "x","y", "x",NA, "x","y", 
                     "x","y", "x","y", "x","y"), 2,7)
 colnames(s.prefs) <- c("A","B","C","D","E","F","G")
 c.prefs <- matrix(c("A","B","C","D","E","F","G", 
                     "A","B","C","D","E",NA,NA), 7,2)
 colnames(c.prefs) <- c("x","y")
 hri(s.prefs=s.prefs, c.prefs=c.prefs, nSlots=c(3,3))
 
## 7 students, 3 colleges with 3 slots each, incomplete preferences:
 hri(nStudents=7, nSlots=c(3,3,3), seed=21, s.range=c(1,3))
 
 s.prefs <- matrix(c('S1', 'S2', NA,
                     'S3', 'S1', NA,
                     'S1', NA, NA,
                      NA, NA,NA,
                     'S2', 'S1', 'S5'),
                   nrow = 3, ncol = 5)
 
 # Note that we explicitly allow for the existence of entries refering to colleges
 # that do not exist. A warning is generated and the entry is ignored.
 colnames(s.prefs) <- c('A', 'B', 'C', 'D', 'E')
 c.prefs <- matrix(c('B', 'C','D', 'A',
                     'C', 'D', NA, NA, 
                     'D', 'B', 'A', 'E'),
                   nrow = 4, ncol = 3)
 colnames(c.prefs) <- c('S1', 'S2', 'S3')
 
 hri(s.prefs=s.prefs, c.prefs=c.prefs, nSlots=c(3,3,3), check_consistency = TRUE)
 
 
## --------------------
## --- Summary plots


## 200 students, 200 colleges with 1 slot each
 res <- hri(nStudents=200, nColleges=200, seed=12)
 plot(res)
 plot(res, energy=TRUE)

matchingMarkets documentation built on Aug. 8, 2023, 5:10 p.m.