In 2015 in Amsterdam, highschool children were allocated to highschools by a `deferred acceptance' algorithm similar to the algorithm for finding a stable matching in a bipartite graph by Gale and Shapley.

http://www.economists.nl/files/20150530-schoolchoice2015may.pdf

The matching procedure based on Gale-Shapley was stopped a few months back:

http://www.nrc.nl/nieuws/2015/10/29/amsterdamse-scholen-stoppen-met-bekritiseerde-matchingsysteem

Then, it was reinstated, with a twist (Gale-Shapley followed by local exchange until pareto optimality):

http://www.nrc.nl/nieuws/2015/11/27/amsterdamse-scholen-gebruiken-toch-bekritiseerde-matching

http://www.verenigingosvo.nl/wp-content/uploads/2013/09/Evaluatie_Matching_simulaties.pdf

In Amsterdam, there is no preference on the side of the schools, only on the side of the students. A stable matching however assumes preference orders both on the side of the schools and on the side of the students. To apply Gale-Shapley, a random preference ordering is fixed for the schools in the above method.

Goals of this bachelor final project:

* Investigate the possibility of using ordinary weighted bipartite matching for this application

* Design and compare suitable weight functions for a bipartite matching formulation, taking the preferences of the students and other criteria into account

* Summarize the advantages and disadvantages of the weighted matching approach relative to the stable matching approach.

* Ideally, both methods are compared with respect to some important aspects on the same dataset.