INVIGILATORS ASSIGNMENT IN PRACTICAL EXAMINATION TIMETABLING PROBLEMS

This paper addresses an invigilator assignment problem. The problem deals with a set of exams, each of which requires a given number of invigilators. The aim is to prepare a conflict-free schedule where all invigilator requirements of the exams are met. In this study, the conflict-free schedule is determined by a mixed-integer linear programming model and a heuristic algorithm that accounts for the following real-life concerns; assignment of the invigilators responsible for an exam, reduction of the number of successive invigilation duties, fair distribution of total workload and unfavorable workload among invigilators and, prioritization of the assignments based on invigilators’ profession. The applicability of the proposed model and the heuristic algorithm has been shown on eight different real-life problems of leading public universities in Turkey and further eight larger-sized examinations set up based on the real settings. In universities, the real schedules are manually prepared by a faculty team. Compared to the assignment of the faculty team responsible for the examination of timetabling in which balancing only the numbers of duties, we achieved to 86% decrease in the total positive error of invigilation hours by fairly distributing the invigilation duties in the model results. Besides, the following improvements are achieved by applying the proposed model; a 56% decrease in the total number of successor assignments, a 44% decrease in the total unfavorable time, and a 23% increase in the total number of department-based assignments. The heuristic algorithm improves the team schedules by 4% in terms of the total positive error of total invigilation hours and 57% in terms of the total number of successive exam assignments. Accordingly, the proposed model and the heuristic algorithm can be used as a decision support tool by the faculty team.