Published in the Asian Journal of Applied Education (AJAE) in 2026, the study developed a Weighted Goal Programming (WGP) model to optimize the even-semester schedule of the Computer Science Study Program at HKBP Nommensen Pematangsiantar University. The findings reveal that the system successfully eliminated all scheduling conflicts involving lecturers and classrooms while improving overall operational efficiency.
Scheduling Challenges in Higher Education
Course scheduling is one of the most complex administrative tasks in higher education. Academic managers must coordinate lecturers, classrooms, laboratories, and time slots while ensuring that all courses can be delivered effectively.
The challenge is particularly significant in study programs that rely heavily on practicum activities. At the Computer Science Study Program of HKBP Nommensen Pematangsiantar University, academic activities must be accommodated using only five regular classrooms and one computer laboratory.
According to information gathered during the study, the scheduling process had previously been performed manually. Preparing a semester schedule required more than a week and frequently resulted in lecturer conflicts, room overlaps, and multiple revisions during the first weeks of implementation.
Yessy Hans Aprilia Manurung noted that these recurring issues highlighted the need for a more systematic and data-driven scheduling approach capable of managing limited resources efficiently.
Using Weighted Goal Programming and Python
To address the problem, the study applied Weighted Goal Programming (WGP) combined with a Mixed Integer Linear Programming (MILP) approach implemented through Python.
In simple terms, WGP allows multiple objectives to be optimized simultaneously by assigning different priority levels to each goal. The system searches for the best possible schedule while respecting essential constraints and minimizing undesirable outcomes.
The model was developed using complete operational data from the 2025/2026 even semester, including:
- 40 active courses
- 161 class sessions
- 34 lecturers
- 5 regular classrooms
- 1 computer laboratory
- 10 daily time slots
- 6 teaching days per week
Several key priorities were incorporated into the optimization process:
- Ensuring all practicum courses receive laboratory access.
- Eliminating lecturer scheduling conflicts.
- Eliminating classroom scheduling conflicts.
- Balancing teaching workloads among lecturers.
- Reducing late-evening classes.
The optimization process was executed using Python's PuLP library and the COIN-OR Branch and Cut (CBC) solver.
Zero Conflicts and Higher Efficiency
The computational model successfully reached a Global Optimal Solution with an objective function value of 19.
The most significant outcomes include:
- Lecturer scheduling conflicts decreased from 3 cases to 0 cases.
- Classroom scheduling conflicts decreased from 2 cases to 0 cases.
- All practicum courses were assigned to the computer laboratory as required.
- All academic activities concluded before 5:10 PM.
- Classroom requirements were reduced from 13 rooms to only 5 regular classrooms and 1 laboratory.
One of the most remarkable findings was the dramatic improvement in facility utilization. While the manually generated schedule effectively required the equivalent of 13 classrooms, the optimized schedule operated successfully using the six facilities already available.
This represents a 61.5 percent reduction in classroom requirements, demonstrating the substantial inefficiencies that can occur in manual scheduling systems.
The model also eliminated late-evening classes, creating a timetable that better aligns with productive learning hours and supports student well-being.
Implications for Universities
The findings suggest that mathematical optimization can serve as a practical solution for higher education institutions facing resource constraints.
For universities with limited classroom space, laboratory facilities, or teaching personnel, similar optimization models could provide several advantages:
- Reduced operational costs.
- Faster schedule preparation.
- Fewer timetable revisions.
- Improved academic service quality.
- Better utilization of existing facilities.
According to Manurung, the study's main contribution lies in integrating laboratory scarcity, lecturer workload balancing, and scheduling preferences into a single optimization framework. This combination remains relatively uncommon in higher education scheduling research in Indonesia.
The model could also be expanded in future studies by incorporating dynamic lecturer availability, particularly for part-time instructors, and by combining WGP with advanced metaheuristic techniques such as genetic algorithms or simulated annealing to handle larger scheduling environments.
Author Profile
Yessy Hans Aprilia Manurung, S.Si. is a researcher affiliated with the Department of Mathematics, Faculty of Mathematics and Natural Sciences, HKBP Nommensen Pematangsiantar University. Her academic interests include operations research, mathematical optimization, linear programming, decision-support systems, and educational scheduling models.
0 Komentar