Development of Assignment Generation System Based on Genetic Algorithm from the Perspective of Classification Development

This study presents an intelligent assignment generation system based on genetic algorithms, designed from the perspective of classification development to provide tailored solutions for diverse student needs. The overall difficulty coefficient of assignments is defined as function F1, the coverage scope as function F2, and the emphasis on key questions as function F3. We developed a comprehensive fitness function that integrates these three metrics, alongside implementing mutation and elite retention strategies to avoid local optima while ensuring robust convergence. All computations were performed using PyCharm. The results indicate that even with a relatively small population size of 50 and only 100 iterations, the system generates high-quality assignments that effectively meet the practical demands of differentiated teaching. Under various parameter settings, the algorithm demonstrates rapid convergence, reaching 82.0% of optimal fitness values within 76 iterations. Analysis of the generated assignments confirms that the system produces stable and reliable outputs, fulfilling real-world application requirements. These findings highlight the potential of genetic algorithms in advancing educational technologies, offering a promising approach for future research and applications.