Generic University Examination Timetabling System with Steepest-Ascent Hill Climbing Hyper-heuristic Algorithm

Abstract

A common problem encountered in various educational institutions is the scheduling of exams, which involves allocating resources such as time and space, as well as considering student schedules within defined hard and soft constraints. This scheduling problem falls under the category of NP-Hard Problems, making it challenging to solve perfectly, and manual scheduling becomes even more difficult. One approach commonly used to address this issue is the Hyper-heuristic method, which generally provides above-average solutions. In this study, the Steepest-Ascent Hill Climbing algorithm is employed to optimize the initial solution generated by the Graph Coloring algorithm. The optimization process, using the Steepest-Ascent Hill Climbing algorithm, was tested over data set from Department of Information Systems, ITS University. The results showed improved timetabling compared to the Simple Hill Climbing algorithm, with lower penalty values. After 10,000 iterations, Steepest-Ascent Hill Climbing achieved an average penalty value of 2.783, whereas Simple Hill Climbing resulted in an average penalty value of 3.943.