Coordinated mobility on demand model for taxi fleet management

Abstract

In the current mainstream industrial practice, Mobility-on-Demand (MoD) systems work in a partially coordinated manner. A centralized driver-allocation system matches a passenger request to the closest idle driver, and the drivers typically follow their own passenger finding strategies. This uncoordinated behavior among drivers leads to a supply-demand imbalance over the operating area. Fleet coordination is a potential solution to mitigate the supply-demand imbalance. In this dissertation, we present our Coordinated Mobility-on-Demand (CMoD) system model for optimal vehicle scheduling and rebalancing. We provide a detailed analysis of the model parameters such as the balancing frequency and the prediction horizon, that are critical for its real-time implementation and assess the performance of the proposed model in a simulated environment. Finally, we present an extension of our model to ensure service fairness for the drivers while operating in this coordinated environment. We quantify service fairness in terms of low disparity of drivers' incomes and high booking success rate and show that the proposed extension of our model leads to a significant improvement in service fairness.