Effect of pre-processors on solution quality of University Course Timetabling Problem

Abstract

This paper describes the effects of pre-processors on the solution quality of the university course timetabling problem. The University Course Timetabling Problem (UCTP) is regarded as both an NP-hard and NP-complete combinatorial optimization problem and is a difficult task since many constraints are needed to be satisfied in order to obtain a feasible solution. The constraints, in this case, are related to the characteristics and regulations of the particular institution. This paper describes the effects of pre-processors which consist of four matrices on the solution quality of the university course timetabling problem as they play a critical role and have significant impact on the solution. In order to evaluate the effects of the pre-processor matrices, the fitness score which is obtained without the pre-processor matrices are tabulated and compared against the fitness score that is obtained with the pre-processor matrices. Seven neighborhood structures are implemented and the experiment is conducted using the datasets obtained from Track 3 of the ITC-2007. The result indicates that there is a significant difference in terms of the fitness score between the solutions obtained without the pre-processors and with the pre-processors, with the latter showing promising results.