Preference-Based Multi-Objective Optimization for Student Transportation: A Machine Learning Approach

This paper presents a novel framework for optimizing student transportation through the integration of the Non-dominated Sorting Genetic Algorithm II (NSGA-II) with Multi-Layer Perceptron (MLP) neural networks. Our approach integrates preference learning with multi-objective optimization to develop personalized mobility solutions that reconcile individual preferences with institutional sustainability goals. Utilizing comprehensive data from over 1,000 students in Strasbourg, we illustrate marked enhancements in transportation efficiency across multiple objectives: cost minimization, comfort maximization, and environmental impact reduction. Our findings indicate that our hybrid methodology outperforms traditional approaches in preference prediction and solution optimization, while maintaining computational efficiency and scalability.