Research on Personalized Training Model of College Vocal Singing Skills Based on Fuzzy Control Algorithm

This paper first identifies three commonly used personalized learning parameters—“learning objectives, learning styles, and cognitive levels”—based on learners' needs. Subsequently, it analyzes the recommendation of personalized learning paths from two dimensions: learners and learning resources, and constructs a personalized learning path recommendation model for vocal singing techniques in higher education institutions. Combining three strategies—division of labor strategy, parameter adaptive adjustment strategy, and simulated annealing strategy incorporating distance factors—to enhance the algorithm's ability to escape local optima, this study proposes a division-of-labor and fuzzy control-based particle swarm optimization algorithm (LDFSPSO) for model solution. Results indicate that the LDFSPSO algorithm can rapidly identify the global optimum of single-peak functions and demonstrates significant advantages when handling multi-peak functions with multiple local extrema points of similar magnitude. The proposed method achieves greater diversity in the resource recommendation sequences for learners and better recommendation effectiveness. Case analysis results also indicate that the LDFSPSO algorithm is more effective in identifying resources that align with learners' needs, and can recommend resources with higher matching degrees for learners with different characteristics.