An Improved Artificial Bee Colony for Course Timetabling

Abstract

The Artificial Bee Colony Algorithm (ABC) is an emerging nature-inspired, metaheuristic optimisation algorithm. In this paper, an improved ABC algorithm is proposed for tackling Curriculum-Based Course Timetabling Problem (CBCTT). The ABC as a population-based algorithm, the initial population is generated using Saturation Degree (SD) followed by Backtracking Algorithm (BA) to ensure that all the solutions in the population are feasible. The improvement loop in ABC used neighbourhood structures severally within the employed and onlooker bees operators in order to navigate the CB-CTT search space tightly. The performance of ABC is tested using dataset prepared by second international timetabling competition (ITC-2007), the ABC is able to achieved good quality results, yet these are not comparable with the best results obtained by other methods. Future work can be directed further improve the ABC operators to achieve a better results.