SELF-ADAPTIVE GENETIC ALGORITHM (SAGA) UNTUK OPTIMASI PENJADWALAN MATA KULIAH: INTEGRASI FUZZY LOGIC DAN COOPERATIVE COEVOLUTION

Hedy Pamungkas
Fandy Neta

Journal of Innovation Research and Knowledge
Vol. 5 No. 2: Juli 2025

University course timetabling in Indonesian higher education represents an NP-hard combinatorial optimization problem with O(n!) complexity affecting 4,500+ institutions. This research develops Self-Adaptive Genetic Algorithm (SAGA) integrating fuzzy logic and cooperative coevolution to address conventional genetic algorithm limitations. SAGA implements cooperative coevolution engine, fuzzy inference system, local search module, and parameter history tracking. Experiments with 1,000 runs on dataset comprising 55 courses, 280+ classes, 44 rooms, and 40 lecturers demonstrate SAGA achieves best fitness 106.850 with 98.9% constraint satisfaction, outperforming Local Search GA by 9.8%. Significant trade-offs including 31-minute execution time and high variability (CV 8.59%) limit practical applications. Algorithm selection framework recommends SAGA for 5% critical cases and Local Search GA for 95% daily operations, demonstrating that algorithmic sophistication does not correlate linearly with practical applicability