10.5281/zenodo.58496
Fadi Safieddine
American University of M. East, Kuwait
Milan Dordevic,
Milan Dordevic
American University of M. East, Kuwait
Ahmed M. Hamza
University of Portsmouth,United Kingdom
Optimization Of Exam Distribution
Zenodo
2016
exam, distribution, EDP, graph-theory, optimization algorithms, cheating
2016-07-22
Journal article
https://zenodo.org/record/58496
Creative Commons Attribution 4.0
Open Access
<p>Academic research findings report level of cheating in exams involving students studying from each other. Several publications have examined means for preventing cheating by means of exam versions, rotations of questions, and addressing social factors. Yet one preventative aspect of exam cheating seems to be neglected and that is exam distribution. In this paper, the authors introduce the Exam Distribution Problem (EDP). Defining a given k versions of an exam in a classroom with n × m chairs, the paper attempts to find the optimal distribution of exam papers such that every two exams of the same version are at maximal distance from each other. Relevant works in Graph-Theory are examined with simulation of Naïve Algorithm (random) and Sequential Release Algorithm (common) for EDP are reviewed. A cost is assigned for instances where two papers of the same version appear in direct proximity thus associated with higher opportunity for cheating. The results showed that the Sequential Release Algorithm did on average no better than the Random Algorithm. Using Optimization Algorithm, the team presents a new approach, the Dichotomous Interleaved Pairing Algorithm (DIP) that achieves minimal adjacency between two identical exam papers and minimal risk of cheating.</p>