Demand management and vehicle routing in dynamic ride-sharing systems

Abstract

Ride-sharing systems contribute to the transition from private cars to a more sustainable utilization of urban traffic infrastructure. To use such systems, travelers indicate their origin and destination via a mobile application, receive a transportation offer, and, if they accept it, are transported to their destination, possibly sharing the vehicle. To ensure long-term acceptance of the service, ride-sharing operators aim to provide a high percentage of travelers with an acceptable transportation offer. To this end, they make use of demand management, i.e., the shaping of demand in its volumes and/or characteristics, and vehicle routing, i.e., routing of the vehicle fleet in order to fulfill transportation requests. In this regard, ride-sharing operators face the challenge that corresponding approaches should ensure both high system performance and fair service conditions, e.g., in terms of traveler fares and driver compensation. The aim of this thesis is therefore to develop soft, i.e. non-monetary, approaches for demand management and vehicle routing in dynamic ride-sharing systems for the benefit of operators, travelers and drivers. To this end, we first focus on implementing the computational framework including state-of-the-art vehicle routing heuristics, as well as on building a comprehensive understanding of controlling demand and its fulfillment in ride-sharing systems. Secondly, we are developing new approaches that focus on demand management and vehicle routing to reduce request cancellations while involving travelers or drivers in the decision-making process. More precisely, we first design heatmaps to support repositioning decisions of idle drivers to balance supply and demand in decentralized ride-sharing systems. Secondly, we design multioptional transportation offers to manage demand while enabling travelers to choose a convenient pickup time from a set of options.