An Elitist-Ant System for Solving the Post-Enrolment Course Timetabling Problem

Ant System algorithms are nature-inspired population-based metaheuristics derived from the field of swarm intelligence. Seemingly, the ant system has a lack of search diversity control since it has only a global pheromone update that intensifies the search. Hence, one or more assistant mechanisms are required to strengthen the search of the ant system. Therefore, we propose, in this study, an elitist-ant system to strike a balance between search diversity and intensification while maintaining the quality of solutions. This process is achieved by employing two diversification and intensification mechanisms to assist both pheromone evaporation and elite pheromone updating, in order to gain a good control over the search exploration and exploitation. The diversification mechanism is employed to avoid early convergence, whilst the intensification mechanism is employed to exploore the neighbors of a solution more effectively. In this paper, we test our algorithm on post-enrolment course timetabling problem. Experimental results show that our algorithm produces good quality solutions and outperforms some results reported in the literature (with regards to Socha’s instances) including other ant system algorithms. Therefore, we can conclude that our elitist-ant system has performed an efficient problem’s specific knowledge exploitation, and an effective guided search exploration to obtain better quality solutions.