Automatic Fleet Balancing in One-Way VSSs via Closed-Loop Dynamic Pricing

Abstract

Vehicle sharing systems (VSSs) are typically divided into two categories, defining distinct trip models. In round-trip VSSs, users must drop off the reserved vehicles at the pick-up station; in one-way VSSs, pick-up and drop-off stations may be different. In round-trip VSSs, the fleet is balanced over the stations by definition, whereas in one-way VSSs, the distribution of the vehicles tends to be unbalanced. Due to this fact, there are situations where no vehicle is available and some reservations must be rejected by the system. Moreover, the VSS organization usually incurs the additional cost of the fleet-balancing operations. Then, even though one-way VSSs are more attractive for the users, they are often not preferred by the service providers. Nevertheless, the scientific literature, which addresses the problem, is quite poor. In this chapter, the problem is addressed from a novel perspective, i.e. it is reformulated as an automatic control problem. The key point is to show that, based on the only (mild) assumption that human decisions are sensitive to changes in the price of the service, a VSS can be accurately modelled as a stochastic dynamical system. The obtained dynamical model is then used to devise a complete feedback control architecture. At the end of the chapter, a thorough simulation campaign shows that the performance of the closed-loop configuration may largely overcome that of the existing VSSs. Such results make this direction of research a promising challenge for both control engineers and people working on intelligent transportation systems (ITSs).