Stackable vs autonomous cars for shared mobility systems: A preliminary performance evaluation

Car sharing is one of the key elements of a Mobility-on-Demand system, but it still suffers from several shortcomings, the most significant of which is the fleet unbalance during the day. What is typically observed in car sharing systems, in fact, is a vehicle shortage in so-called hot spots (i.e., areas with high demand) and vehicle accumulation in cold spots, due to the patterns in people flows during the day. In this work, we overview the main approaches to vehicle redistribution based on the type of vehicles the car sharing fleet is composed of, and we evaluate their performance using a realistic car sharing demand derived for a suburban area around Lyon, France. The main result of this paper is that stackable vehicles can achieve a relocation performance close to that of autonomous vehicles, significantly improving over the no-relocation approach and over traditional relocation with standard cars.