COMBINATION OF GENETIC ALGORITHM AND TABU LIST IN THE CASE OF EXAM SCHEDULING

Abstract

The exam scheduling system aims to provide solutions for scheduling exams that have been done manually. Several studies on scheduling have been done, one with a genetic algorithm. But there are still weaknesses that occur in the genetic algorithm is often trapped in local optimum. This study developed a scheduling system using a combination of genetic algorithms and tabu lists to address the weaknesses of the previous system. This research uses Immanuel Christian University data on Informatics Engineering course. Genetic algorithm and taboo list results show that genetic algorithm succeeded in avoiding local optimum by using taboo list as a requirement that genetic result value should not be used if the result is already in taboo list. The process of scheduling using genetic algorithms and tabu lists can reduce the generation of more than the genetic algorithm. But the disadvantage of genetic algorithms with taboo lists is that process time tends to be slower than the genetic algorithm without taboo lists.