Modélisation et évaluation d’un système de covoiturage dynamique optimisé

Abstract

Dynamic ridesharing has become a promising solution to facilitate the matching of passengers and drivers, optimizing vehicle use and reducing environmental impact. However, to ensure the success of these systems, it is essential to establish effective matches between passengers and drivers. This is the context of our work. Our main objective is to use the "Firefly" metaheuristic algorithm to solve the dynamic matching problem between drivers and passengers. The solution we propose in this thesis takes into account spatio-temporal constraints, capacity constraints and waiting time constraints. Our model aims to optimize matching by minimizing both passenger waiting times and total distance traveled. To evaluate the effectiveness of our approach, we conducted a study using data generated from a real geographical area located in the wilaya of Guelma. This rigorous evaluation enabled us to analyze the performance of our algorithm in detail, and to highlight its advantages and limitations. The results obtained show that the Firefly algorithm achieved more efficient matches, reducing both passenger waiting times and distances travelled. These results underline the effectiveness of our approach in improving the quality and efficiency of ridesharing systems.