Rawlsian assignments

This code can be used to compute the Rawlsian assignment from a given one-sided matching instance using the algorithm as described in the accompanying paper by Demeulemeester & Pereyra (2023).

Initialization

The code uses the Gurobi optimization software (version 9.1.1), and calls it from a Jupyter Notebook. Gurobi offers a free academic license.

Use

The one-sided matching instance for which you want to compute the Rawlsian assignment can be read from a .csv file. To read in a data file, there are two options:

• rank: each column of the file represents an object, and the corresponding number indicated the position of the object in the preference list of an agent. Assuming three objects $a,b,c$, preference list (3,1,2) refers in this setting to object $b$ being most preferred, then object $c$, and object $a$ being least preferred.
• list: the objects are listed from most preferred to least preferred. Preference list (3,1,2) refers to object $c$ being most preferred, then object $a$, and object $b$ being least preferred.

A simple working example to compute and to visualize the Rawlsian assignment is:

file_name = "Example_SP_axioms.csv"
MyData = Data(file_name, "list")
Sol = Rawls(MyData)
Sol.visualize()