Faculty of Computing

Search Results

Now showing 1 - 1 of 1

Embargo
Evaluating Genetic, Reinforcement Learning, and Colony Optimization Algorithms for Scalable University Timetable Scheduling
(Institute of Electrical and Electronics Engineers Inc., 2025-07-03) Weerasinghe K.D.E.I.; Wijayawardhana G.L.C.N.D.; Udayantha D.M.S.; De Silva K.H.P.N.; Perera, J; Hathurusinghe, S
University timetable scheduling, a complex optimization task, often lacks comparative algorithmic analysis under standardized, resource-scarce conditions. This research introduces an integrated framework to benchmark Genetic Algorithms (GA), Reinforcement Learning (RL), and Colony Optimization (CO) under identical, extreme resource scarcity (4 rooms for 195 activities), a key differentiator. Using standardized metrics, results revealed distinct trade-offs: GAs offer rapid initial solutions but lack flexibility; RL shows efficient resource use but requires extensive training; CO demonstrates adaptability. All algorithms struggled to assign every activity due to imposed resource limits, highlighting performance under severe contention. The developed platform, with configurable scoring and role-based access, is a practical contribution, offering data-driven insights for algorithm selection. This study offers systematic three-way comparison of these techniques under uniform, high-pressure criteria, informing future hybrid solutions.