NEW EXAMINATION TIMETABLING ALGORITHM USING THE SUPERSTAR ASSIGNMENT TECHNIQUE

In this paper, a new examination timetabling algorithm, SAT, is introduced. In order to solve the current problems that SAT algorithm can meet the requirement under the limited of all related resources and factors. It combined with rules, constraints, exceptions and priorities. This algorithm works in three steps: pre-processing, creating superstars and getting rid of superstars. SAT mechanisms sort all related parameters and factors, then determine stars and superstars of each related parameter. Each iteration of algorithm is trying to assign all superstars of all parameters as targets for processing. As well as, the process mechanism is run for putting the output into a suitable timeslot. To prove the algorithm implementation, a dataset from semester 1/2016 of Registration Centre, Ramkhamhaeng University has been selected as test subject. This dataset consists of 20/2 days/periods, 85,000 registered students, 1,325 subjects, 813,253 seats, and 11/76/22,582 buildings/rooms/seats per day. The proposed model could be solved the current problems and shown details of subject such as the colour of answer sheet, room and so on. It could be done within less than two hours, meanwhile the current system took at least one month. For the future work, the large scale of algorithm is to be improved and developed into more dynamic version for a larger volume of data and handling more complicated constraints.