Comparison of different delivery models using self-driving robots

Last-mile logistics has gained increasing attention in the last few years. Driven by the rise of e-commerce and the growing demand for fast deliveries, companies are exploring several strategies to optimize last-mile operations. One of the most promising options in city logistics is the truck-and-robot concept that involves a truck transporting parcels and autonomous robots. This delivery system exploits a combination of drop-off points and robot depots. When a truck arrives at a drop-off point, robots are released directly from the truck to autonomously deliver goods. On the contrary, if a depot is reached, parcels are loaded on the robots, that are stored and charged in that location, so that they can be delivered from the depot. In this thesis, three delivery models have been simulated and compared in order to determine the most effective approach. The compared methodologies are: the previous truck-and-robot method, which enables parallelized deliveries, a standard approach, where the driver directly visits customers, and a mixed strategy where a portion of clients is served by robots and the other one by trucks. The obtained results on literature’s instances show that the collaboration between trucks and robots is more convenient than the traditional truck service. Therefore, self-driving robots are particularly suitable for simultaneous deliveries as they reduce times and costs, respecting time windows constraints. Nevertheless, traditional logistics remain faster and more efficient in scenarios where parallelization is not feasible.