Academic Timetable Generation Using Abandoned and Reborn Solution Mechanism of Particle Swarm Optimization
Proceeding: The Fourth International Conference on Digital Information Processing and Communications (ICDIPC)Publication Date: 2014-03-18
Authors : Paulus Mudjihartono;
Page : 145-151
Keywords : timetable; abandoned-reborn mechanism; problem mapping; effectiveness; objective value;
Abstract
One of the challenging heuristic problems is that how to generate an academic timetable. Many searching methods have already been applied to take care of the problem but yet some drawbacks exist. The drawbacks are commonly about the convergence, the speed, and the effectiveness of the algorithm. This paper aims to apply modified Particle Swarm Optimization (PSO) to give a better solution rather than that in the ordinary PSO. The representation of the academic timetable is implemented by making a simple list of the sessionclassrooms. A list of the session-classrooms describes one academic timetable solution. The solution may be better than the others depending how good it avoids the constraint violations. In modified PSO, some solutions will not be accepted if their penalties are too large. The algorithm simply abandons them and recreates some new solutions instead of correcting them. We used the objective value to justify the algorithm effectiveness. With the same number of epoch 100, 20 particles, 88 lectures, 35 constraints, we found that ordinary PSO yielded -37.3 and abandoned and reborn mechanism yielded -26.2. Both algorithms give the solution with minor constraint violations, but the modified PSO shows 29.3% higher in the objective value.
Other Latest Articles
- Monitoring Human Blood Pressure for U-Healthcare Using ISO/IEEE PHD Standard
- A Design of U-Health System on Smart Phone Using ISO/IEEE 11073 PHD Standard
- Degree of Diffusion in Real Complex Networks
- Construct a Mobile Device Enabled Ubiquitous E-learning Environment in Cloud
- Indoor Next Location Prediction with Wi-Fi
Last modified: 2014-03-24 23:06:32