Relocation in One-Way Station-Based Car Sharing Systems: Conventional Versus Partly Autonomous Vehicles

Abstract

Our aim with this paper is to formulate a platoon relocation model and to apply it on a large-scale one-way car sharing system. We analyze the system performance in terms of realized customer trips and relocation costs with regard to conventional, platoon or none relocation tours. We analyze the efficiency potential of platoon relocation on simplified instances that mimic demand and city structures. In these examples, the platoon approach can be up to 60% more efficient in relocating cars compared to the conventional one. We formulate a time-expanded network model for platoon based relocation in station-based one-way car sharing systems and apply it to a self-designed case study from the city of Hanover. In order to cope with runtime restrictions, we delete redundant network arcs. Our results show that in the large-scale Hanover case example, the platoon tours lead to a moderate increase of customer trips and reduced relocation costs.