Abstract:
The university course timetabling problem (UCTP) schedules a set of events into timeslots and suitable rooms under various constraints according to the student enrollment such that the possibility of allocations is maximized. It is solved by introducing an innovative idea based on students groupings and then using ant colony optimization (ACO) algorithm. Disjoint groups of students, based on selected events, are created in such a manner that a student must belong to exactly one group. This procedure is executed by excluding the student from the further selection, once it is selected in a group. Events obtained after this grouping of students are assigned to timeslots and rooms by applying the ACO algorithm. The proposed algorithm is tested on a number of benchmark UCTP instances. The fitness function value is used as a performance measure. Each problem instance is run independently for several times and the solution among them with the least value of fitness function is selected as the solution of that problem instance. It is found that the proposed algorithm is computationally efficient when compared with other existing algorithms.