Adaptive Web Course Scheduling via Multi-Armed Bandit Selection of ACO and PSO Algorithms

The scheduling of courses in higher education is a complicated combinatorial optimization issue that has to meet many institutional constraints and balance the preferences of lecturers and the availability of resources. Ant Colony Optimization (ACO) and Particle Swarm Optimization (PSO) are metaheuristic algorithms that have been extensively used to solve this problem, but their performance frequently changes in response to the nature of the scheduling situation. This paper proposes a dynamic scheduling model based on a Multi-Armed Bandit (MAB) mechanism that is used to dynamically choose between ACO and PSO in the optimization process. The suggested solution will enhance the strength of the algorithms by integrating the strengths of the two algorithms without compromising on the computational efficiency. The framework is applied in a web-based scheduling system which facilitates interactive data management, constraint setup, and lecturer-specific scheduling requests. Real data of faculty scheduling involving sixteen study programs, almost a hundred courses per program, and forty-one lecturers were used as experimental data. The evaluation involves quantitative experiments and qualitative validation of the lecturer surveys. Findings indicate that the adaptive framework can provide a balanced trade-off between the quality of solutions and the performance of the runtime with positive user feedback in the aspects of usability, efficiency, and the acceptance of the system. These results reveal the possibility of the suggested system to be implemented in real-life academic scheduling settings.