Perbandingan Genetic Algorithm dan Queue-Based Scheduling untuk Penjadwalan Kuliah Otomatis di Perguruan Tinggi

Pratama, Muhammad Demas Adi
Kurniawan, Wakhid

TIN: TERAPAN INFORMATIKA NUSANTARA
Vol 6 No 11 (2026): April 2026

Course scheduling is a fundamental activity in academic management; however, the complexity of room and teaching constraints often triggers schedule conflicts when processed conventionally. This study aims to design an automated course scheduling system at Universitas Muhammadiyah Karanganyar (UMUKA) by comparing the effectiveness of the Genetic Algorithm (GA) and Queue-Based Scheduling (QBS). The system was developed using the Laravel framework and a NoSQL Database. Comparative experiments were conducted through six dynamic scenarios (ranging from 20 to 263 courses), measured using three metrics: the final number of conflicts, fitness value, and computational time. The results indicated that QBS was absolutely superior in computational speed (maximum 4.18 seconds) but failed to produce proportional schedule quality, as the fitness value remained at 0. Conversely, GA successfully generated conflict-free schedules in medium workloads and compensated for minor conflicts in massive workloads by achieving a significant spike in fitness value (reaching 462.2). Although requiring relatively longer computational time (36.2 seconds at the highest load), this duration remains highly efficient. In conclusion, GA is recommended as the primary algorithm for course scheduling systems due to its ability to perform global optimization proportionally. The main contribution of this research is providing empirical foundations and a comparative architecture for higher education institutions in selecting the most adaptive scheduling algorithm for local room infrastructure limitations and hard-constraint complexity.