Analyzing the effects of different problem settings on the satisfaction levels of students and supervisors in the student-supervisor allocation problem

Abstract

Appropriate student-supervisor allocations are critical for the effective management of graduate studies as well as the overall success of graduate programs. In this study, as shown in Figure A, three different problem settings are considered for the student-supervisor allocation (SSA) problem where, in addition to taking into account student and supervisor preferences, (i) supervisor capacities are not considered, (ii) individual supervisor demands are considered in terms of their capacities and (iii) average supervisor workload is considered in terms of their capacities. Mathematical programming models and a heuristic algorithm are proposed for the solution of each problem setting. In order to test the validity of the proposed approach, two different datasets, a small- and a large-scale one, are generated. Using these datasets, the proposed mathematical programs and the heuristic algorithm are solved for different weighting strategies of students' and supervisors' preferences. The results of the mathematical programming models and the heuristic algorithm are compared. The SSA results for different problem settings and allocation strategies are analyzed according to performance measures defined in terms of the satisfaction levels of students and supervisors.