Webinar by Can Akkan (Professor, Sabancı University)

11/05/2020 18:00

"A Bi-Criteria Simulated Annealing Algorithm for the Robust University Course Timetabling Problem Subject to Period-Unavailability Disruptions"

A bi-criteria version of the curriculum-based course timetabling problem of ITC-2007 (International Timetabling Competition) is solved to identify an approximation to the optimal Pareto front. The two criteria are the penalty function as defined in ITC-2007 and a robustness function. Robustness is needed to deal with disruptions that occur in the form of a feasible period that is used by a course becoming infeasible for that course. In such situations, the timetable has to be updated, in essence re-optimized, while ensuring that the changes are not excessive. A timetable is said to be robust if, when disrupted, it can be re-optimized without significantly lowering its quality in terms of the objective function and keeping it relatively stable. We define a robustness function that takes into account both the solution quality and stability concerns. We define the problem with the possibility of multiple disruptions as a multi-objective stochastic combinatorial optimization (MOSCO) problem. We propose a Multi-objective Simulated Annealing (MOSA) algorithm that makes use of a Monte Carlo sampling approach to solve the problem. Performance of the algorithm is tested on the ITC-2007 instances.

Please register at: https://lnkd.in/dWNkbFm