In the Mobility-on-Demand (MoD) systems such as Uber, Hubway, etc., vehicles transport customers within an urban environment. In a highly variable environment, when some origins and destinations are more popular than others, the systems inevitably become out of balance: vehicles accumulate in some neighborhoods and become depleted at others. Over the last decade, researchers have proposed models that rebalance the distribution of vehicles in a cost-effective manner, while also maximizing the throughput of the system.
In this talk, we will discuss some problems that arise while designing such rebalancing models. We will determine the service level requirement bounds at different places using queuing theoretic models. We will compare the domain-specific routing problems that arise in applications such as bike-inventory rebalancing and autonomous taxi vehicle rebalancing. Lastly, we will discuss the role of dynamic pricing in an efficient matching of supply with demand in a two-sided market like Uber, and conclude by highlighting some open problems in the area.
Prof. Evimaria Terzi, Prof. John Byers, Prof. Mark Crovella, Prof. George Kollios.