Construction of Initial Solution Population for Curriculum-Based Course Timetabling using Combination of Graph Heuristics

Abstract

The construction of population of initial solution is a crucial task in population-based metaheuristic approach for solving curriculum-based university course timetabling problem because it can affect the convergence speed and also the quality of the final solution. This paper presents an exploration on combination of graph heuristics in construction approach in curriculum based course timetabling problem to produce a population of initial solutions. The graph heuristics were set as single and combination of two heuristics. In addition, several ways of assigning courses into room and timeslot are implemented. All settings of heuristics are then tested on the same curriculum based course timetabling problem instances and are compared with each other in terms of number of population produced. The result shows that combination of largest degree followed by saturation degree heuristic produce the highest number of population of initial solutions. The results from this study can be used in the improvement phase of algorithm that uses population of initial solutions