Analysis of the Impact of Population Size on Computational Time Efficiency in Genetic Algorithms for Course Scheduling

Course scheduling is a complex problem within higher education systems that requires automated and efficient solutions. Genetic algorithms are widely employed due to their capability to explore large solution spaces. However, the efficiency of the algorithm highly depends on key parameters, one of which is the population size. This study aims to evaluate the impact of population size variation on computational time efficiency in the implementation of genetic algorithms for course scheduling. The study was conducted through MATLAB-based simulations in a real academic environment, testing population sizes ranging from 20 to 1000. Other parameters were held constant to ensure that execution time was influenced solely by changes in population size. The results reveal a non-linear relationship between population size and computational time, with the highest efficiency achieved at a population size of 300–400 individuals, yielding an average execution time of approximately 0.4924 seconds. Extremely small or large population sizes were shown to produce suboptimal execution times. These findings highlight the importance of empirical evaluation in algorithm parameter selection, particularly in systems with processing time constraints, such as course scheduling.