Abstract:
This study presents an automated timetabling system developed to address the inefficiencies
of manual class scheduling at the University of the Philippines Manila
(UPM), where diverse departmental constraints often result in conflicts and delays.
The web-based system integrates a Genetic Algorithm (GA) with two local search
optimization techniques—Great Deluge Algorithm (GDA) and Simulated Annealing
(SA)—to construct and refine timetables that satisfy hard constraints (e.g., room
capacity, faculty availability, conflict-free schedules) and optimize soft preferences
(e.g., preferred rooms, accessibility). Users can upload CSV data, visualize schedules
through dynamic calendar views, and export outputs, while a utilization scoring
feature enables assessment of room efficiency. Built on a Django backend, the platform
significantly reduces scheduling time and errors, demonstrating strong performance
across complex scenarios and offering adaptability for broader academic use.
By combining global and local search strategies, the system not only delivers highquality
timetables tailored to institutional needs but also provides a scalable model
for future integration in the university system.
