International Journal of Scientific Research and Engineering Development

This paper presents a simulation-based machine learning approach to predict elevator congestion in university buildings. A discrete-event simulation model was designed using multiple operational factors such as arrival rate, waiting area capacity, queue length, peak hours, and environmental conditions (day type). The simulation generated a synthetic dataset of 10,000 samples, which was used to train two supervised learning models: Logistic Regression and Random Forest. The performance of the models was evaluated using confusion matrices and four standard metrics: accuracy, precision, recall, and F1-score. While Random Forest achieved slightly better performance than Logistic Regression, both models showed limited recall for congestion cases due to class imbalance. The proposed framework demonstrates how SimPy-based simulation and machine learning can support data-driven planning for smart campus mobility and elevator management.