This paper investigates solving university timetabling problems using convex programming. Historically, such problems are solved using mixed integer linear programming or using heuristic approaches that require considerable computation and the solutions are suboptimal. The new formulation uses the L1 norm penalty that promotes solution sparsity knowing that the vector of decision variables is essentially sparse. We test the new formulation on the international exam timetabling benchmark problem and demonstrate the efficiency of the technique compared to mixed integer linear programming.
Track: Operations Research
Published in: 1st GCC International Conference on Industrial Engineering and Operations Management, Riyadh, Saudi Arabia
Publisher: IEOM Society International
Date of Conference: November 26
-28
, 2019
ISBN: 978-1-5323-5951-4
ISSN/E-ISSN: 2169-8767