Final exams scheduling for univeristies. (c2001)

LAUR Repository

Show simple item record

dc.contributor.author Timany, Mazen Nuhad
dc.date.accessioned 2011-10-14T12:09:18Z
dc.date.available 2011-10-14T12:09:18Z
dc.date.copyright 2001 en_US
dc.date.issued 2011-10-14
dc.date.submitted 2001-07
dc.identifier.uri http://hdl.handle.net/10725/777
dc.description Includes bibliographical references (p. 45-47). en_US
dc.description.abstract Scheduling final exams for large numbers of courses and students in universities, such as the Lebanese American University (LAU), is an intractable problem. In order to solve this problem, the approach must be efficient, flexible and adaptable. Conflicts occur when multiple exams are scheduled for the same student at the same period (simultaneously), and unfairness to a student refers to consecutive exams (two exams directly after each other) or more than two exams on the same day (referred to as mUltiples). A good exam schedule would aim for minimizing conflicts and the two unfairness factors based on user-assigned weights associated to these three factors and subject them to some constraints. Likewise, since a limited number of rooms are available in each exam period, an additional constraint concerned with room violations is added to achieve the goal of minimizing room violations. All constraints may be violated if necessary, since it is almost impossible in real world situations to find a solution without violating any constraint. In this work, we first formulate the problem as a modified weighted-graph coloring problem and adapt two natural optimization algorithms: Simulated Annealing and Genetic Algorithm; in addition to a clustering based algorithm (FESP), and a hybrid of natural optimization and clustering based algorithms (FESPSA) for solving the exam scheduling problem taking into account the specific objectives and constraints of LAD. Then, we compare these algorithms with each other as well as with the manual procedure done by the registrar's office. The comparison is done using realistic data taken from LAU for six semesters. Our experimental results show that simulated annealing gives better exam schedules than genetic algorithms, FESPSA, FESP and manual scheduling. All algorithms were run on different exam days ranging from five to ten. Simulated Annealing stayed to show the best results in all semesters in all days variations. Moreover, Simulated Annealing shares with the Genetic Algorithm more flexibility to accommodate various user constraints. On the other hand, FESPSA showed better results in terms of conflicts and unfairness factors among all exam days when compared with FESP. Moreover, FESPSA also proved to be better than FESP when dealing with room violations. That is, FESPSA minimized the number room violations much better than FESP. en_US
dc.language.iso en en_US
dc.subject Simulated annealing (Mathematics) en_US
dc.subject Scheduling -- Computer programs en_US
dc.title Final exams scheduling for univeristies. (c2001) en_US
dc.type Thesis en_US
dc.term.submitted Summer I en_US
dc.author.degree MS in Computer Science en_US
dc.author.school Arts and Sciences en_US
dc.author.idnumber 199110680 en_US
dc.author.commembers Dr. Ramzi Haraty
dc.author.commembers Dr. George Nasr
dc.author.woa RA en_US
dc.description.physdesc 1 bound copy: xiii, 47 leaves; ill., tables; 30 cm. available at RNL. en_US
dc.author.division Computer Science en_US
dc.author.advisor Dr. Nash'at Mansour
dc.keywords Exam scheduling en_US
dc.keywords Timetabling en_US
dc.keywords Constraint programming en_US
dc.keywords Simulated annealing en_US
dc.keywords Genetic algorithms en_US
dc.keywords Optimization heuristics en_US
dc.identifier.doi https://doi.org/10.26756/th.2001.16 en_US
dc.publisher.institution Lebanese American University en_US

Files in this item

This item appears in the following Collection(s)

Show simple item record

Search LAUR

Advanced Search


My Account