Design and statistical analysis of initial solution construction approach in curriculum based course timetabling problem

• Main Authors: Wahid, Juliana, Mohd Hussin, Naimah
• Format: Conference or Workshop Item
• Language: English
• Published: 2017
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://repo.uum.edu.my/24481/1/SICONSEM%202017%204%206.pdf; http://repo.uum.edu.my/24481/

This paper investigates the construction phase approach in which the sequential order of the courses/lectures applied several combination of graph heuristics to generate a population of initial solutions in curriculum-based course timetabling (CBCTT).The construction of population of initial solution is a prerequisite in a population-based metaheuristic implementation. To produce a population of initial solution require algorithm that can produce multiple feasible solutions and these solutions must be diverse. This process is a crucial task because it can affect the convergence speed and also the quality of the final solution (Rahnamayan, Tizhoosh, & Salama, 2007).This study able to produce a set of initial solution, therefore it is able to contribute to the improvement phase of approach that uses population of initial solutions such as ant colony optimization (ACO) (Socha, Joshua, & Michael, 2002), genetic algorithm (GA) (Lewis & Paechter, 2005), and harmony search algorithm (HSA) (Al-Betar & Khader, 2010).The approach in this study also shows that a feasible timetable can be found for numerous data set problems.