Thanawat Unprasertporn. A hybrid multi-objective genetic-based discrete particle swarm optimization algorithm with a local search for solving the post enrolment based course timetabling problem. Master's Degree(Mathematics with Computer Science). King Mongkut's University of Technology North Bangkok. Central Library. : King Mongkut's University of Technology North Bangkok, 2019.
A hybrid multi-objective genetic-based discrete particle swarm optimization algorithm with a local search for solving the post enrolment based course timetabling problem
Abstract:
The Post Enrolment Based Course Timetabling Problem (PECTP) is a part of university course timetabling problem (UCTP), which is the problem that occurs continuously in all universities. It is about the allocation of subjects, so called events, into time slots and suitable rooms according to students enrolment. In addition, arranging the optimal and effective timetable is a quite hard task due to the complexity of the problem itself. The solution of this problem must satisfy all hard constraints and meets the soft constraints as much as possible. This problem is therefore classified as a combinatorial optimization problem, with the difficult level of NP-Complete problem resulting in the highly time-consuming of solving the problem. Moreover, the timetabling problem in real world situation was considered as the multiple objective problem more than a single objective problem. Continuously, in this research, we have improved the representation of a solution of genetic-based particle swarm optimization (GDPSO) for better results because the standard particle swarm optimization (PSO) was designed to work with the continuous nonlinear functions more than a discrete function. We also have developed a multi-objective GDPSO combined with local search approaches for solving the PECTP, and it will be tested with Metaheuristics Network (MN) instances. Last, the experimental results of the proposed hybrid approach will be compared with other algorithms from the literature in terms of the quality of the solutions rather than the computing time of getting the solutions.