Incentive-Based Rebalancing of Bike-Sharing Systems

This paper proposes an incentive-based approach for rebalancing bike-sharing systems where customers are offered discount to pick up bikes from nearby stations that are expected to become full in the near future. The main contribution of this work is twofold: (1) we develop a customized station object in the Simio simulation software to facilitate modeling of bike-sharing systems and reduce the burden on the modeler by eliminating the need to code the basic functionalities of a bike station; and, (2) we develop a discrete event simulation model of a real-world bike-sharing system (CitiBike) using instances of the customized station object to evaluate the effectiveness of pickup incentives in rebalancing the system. The model is calibrated using historic data and the results confirm the effectiveness of such incentive-based rebalancing scheme. More specifically, the results suggest that while incentives help improve bike availability in general throughout the system (i.e., better balance and service), offering too many incentives can in fact reduce total profit due to decreased marginal profit per ride.