These data files correspond to the experiments reported in Section 3 of the following paper:

I. McBride and D.F. Manlove.  An Integer Programming Model for the Hospitals/Residents Problem with Couples.  In Proceedings of OR 2013: the International Conference on Operations Research, pages 293-299, Springer, 2014.

Publisher final version: http://dx.doi.org/10.1007/978-3-319-07001-8_40
Author final version: http://eprints.gla.ac.uk/98265/

The data in folder "Varying Number of Residents" refer to the experiment described on Page 296 (results reported in Figure 1 on Page 297).

The data in folder "Varying Percentage Couples" refer to the experiment described on Page 297 (results reported in Figure 2 on Page 297).

Each instance file name is in the following format:

<number of residents> - <number of hospitals> - <number of couples> - <number of posts> - <min resident pref list length> - <max resident pref list length> - <posts distributed evenly?> - <resident popularity> - <hospital popularity> - <Iteration number>.txt

Each instance file contains the following data:

<number of residents n_1>
<number of hospitals n_2>
<number of couples x>
<number of posts>
<min resident pref list length>
<max resident pref list length>
<posts distributed evenly?>
<resident popularity>
<hospital popularity>

<2x preference lists for the residents who are involved in a couple>
<n_1-2x preference lists for the single residents>

<n_2 hospital preference lists>

Resident and hospital IDs count from 0.

Residents 0,1,...,2x-1 are the residents who are involved in a couple, and the couples comprise resident pairs (2i,2i+1) (0<=i<=x-1).  The preference lists of the residents involved in couple (2i,2i+1) are in the following format:

<resident ID 2i> <hospital ID a0> <hospital ID a1> ... <hospital ID ak>
<resident ID 2i+1> <hospital ID b0> <hospital ID b1> ... <hospital ID bk>

This means that the joint preference list for the couple (2i,2i+1) is then:

<first-choice hospital pair (a0,b0)> <second-choice hospital pair (a1,b1)> ... <last-choice hospital pair (ak,bk)>

The preference lists of the single residents are in the following format:

<resident ID> <first-choice hospital ID> <second-choice hospital ID> <third-choice hospital ID> ...

The preference lists of the hospitals are in the following format:
 
<hospital ID> <hospital capacity> <first-choice resident ID> <second-choice resident ID> <third-choice resident ID> ...

Note that for the experiment "Varying Number of Residents", the number of residents increases in increments of 50, not 30 as reported in the paper.

Also note that due to data protection legislation, data from the Scottish Foundation Allocation Scheme, which was used to produce the results shown in Table 1 on Page 298, cannot be made publicly available.