Integrating Artificial Intelligence and Optimization Techniques for Efficient Delivery

The rise of e-commerce and its significant impact on transport and logistics have prompted companies to question the efficiency of parcel deliveries. The question of parcel delivery efficiency comes up against a well-known classic problem in logistics, namely the vehicle routing problem (VRP), which consists in optimizing vehicle route predictions in order to save time while maximizing delivered customer satisfaction. This is the objective of this article, which proposes to combine machine learning techniques linked to artificial intelligence, such as K-Means, DBSCAN and OPTICS, to group delivery points into clusters, thus reducing the complexity of the problem. Building on these clustering methods, our paper highlights the use of the Tabu search algorithm to optimize delivery rounds within each cluster. By exploring different optimization approaches, we plan to deploy the Tabu search algorithm on several clusters while adding a constraint on the number of deliveries per round, which will result in more efficient delivery round solutions tailored to our operational constraints. Using this combined approach, our paper proposes practical recommendations to help decision-makers improve their delivery service and respond effectively to the logistical challenges of the e-commerce sector.