A greedy gradient-simulated annealing hyper-heuristic for a curriculum-based course timetabling problem

The course timetabling problem is a well known constraint optimization problem which has been of interest to researchers as well as practitioners. Due to the NP-hard nature of the problem, the traditional exact approaches might fail to find a solution even for a given instance. Hyper-heuristics which search the space of heuristics for high quality solutions are alternative methods that have been increasingly used in solving such problems. In this study, a curriculum based course timetabling problem at Yeditepe University is described. An improvement oriented heuristic selection strategy combined with a simulated annealing move acceptance as a hyper-heuristic utilizing a set of low level constraint oriented neighbourhood heuristics is investigated for solving this problem. The proposed hyper-heuristic was initially developed to handle a variety of problems in a particular domain with different properties considering the nature of the low level heuristics. On the other hand, a goal of hyper-heuristic development is to build methods which are general. Hence, the proposed hyper-heuristic is applied to six other problem domains and its performance is compared to different state-of-the-art hyper-heuristics to test its level of generality. The empirical results show that the proposed method is sufficiently general and powerful.