Estelito Ii Y. Buenavista, Matthew Cedric D. Calaycay, Christine F. Peña
University course timetabling (UCTP) is an NP-hard combinatorial optimization problem balancing hard feasibility constraints and soft penalties. This paper presents Werepire, a sequential hybrid that pairs the Grey Wolf Optimizer (GWO) for diversified global search with a Bat Algorithm (BA) stage for local refinement. A UniTime-based initializer provides diverse feasible seeds, and a shared ITC 2019 validator measures solution quality. On three representative instances (muni-fi-spr16, muni-fsps-spr17, tg-fal17), BA remained the strongest baseline for municipal-scale cases, while Werepire achieved the best penalties on the constraint-dense tg-fal17 with competitive runtime. Friedman-Nemenyi analysis indicated that cross-instance rank differences were instance-specific: BA and Werepire formed a tie group overall yet alternated leadership per dataset, whereas GWO consistently trailed. These results suggest that GWO-to-BA handoffs mitigate premature convergence on difficult curricula while incurring limited overhead, whereas BA remains a pragmatic default elsewhere. © 2026 IEEE.
University of San Carlos, Dept. of Computer, Information Sciences, and Mathematics, Cebu City, Philippines
