Impact of relocation strategies for a fleet of shared automated vehicles on service efficiency, effectiveness and externalities

The introduction of taxi-like transit services operated by shared automated vehicles comes into sight with the development of vehicle automation. In this paper, the operation of such a service is simulated for a generic grid network in order to determine the impact of different relocation strategies for idle vehicles on passenger waiting time, empty mileage and parking needs. The tested strategies consist of remaining idle at the latest drop-off location, returning to the initial position, relocating to a random location, relocating according to anticipated demand or relocating to a zone with a low vehicle supply. For the simulated case study, remaining idle outperformed the other relocation strategies in terms of service efficiency and service effectiveness, while the strategy of evenly or randomly dispersing vehicles over the network lead to largest reduction of the number of parked vehicles per link, and the strategy of anticipating demand to largest reduction of deadheading mileages.