Abstract:

Many recent researches in university timetabling require huge amount of computation power to find feasible solutions and to make matter worse, these computation complexity increases exponentially with problem size. As universities are continuously growing, in terms of size, requirement and operation, it may actually slow down the overall study performance in schedule planning, especially when in cases where requirements are changing rapidly. UniT-CORECTS, the proposed new system, will become a useful tool to help schedule planners detect and rectify conflicts visually. UniT-CORECTS is specially designed and developed to check if conflicts exist within planned timetable. The planned timetable can be obtained as a result of manual study performed by timetable planners or from a fully automated timetabling system. When conflicts are detected, it allows users to rectify the problems manually, with   the aid   of presenting the problems in graph form and allowing simple stroke gestures be used to make the corrections easily. UniT-CORECTS will apply a hybrid method using a modified graph annealing algorithm and the proposed STAR-quadrants to display the conflict detection of the entire timetable. Simple stroke gesture is used to expedite the conflict resolution process. UniT-CORECTS is aimed to ease the study of the user, not just by making timetable planning more interesting, but also making some tedious configuration study faster and thus improves overall study performance efficiency. Besides applying the technique to solve existing course timetabling problems, it is expected that the approach can be refined to study well in conflict detection and rectification in transportation schedules. However, future research has to be concentrated on improving the solution to include the visibility of conflicts caused by soft constraints alongside the hard constraints.