CrowdDeliver: Making Citywide Packages Arrive as soon as Possible

Abstract

It is still difficult to make the express service profitable, despite recent years have witnessed the great demand on and attempts at the service of package express shipping. The main barrier may be due to that the speedy usually implies a higher sending frequency. To strike a trade-off between the two conflicting objectives, we propose a new idea that exploits the existing taxi mobility to transport packages collectively (i.e., the relays among different passenger-occupied taxis), without hurting the service quality to passengers too much. In more detail, we propose and develop a novel framework called CrowdDeliver, which is a two-phase approach to plan package delivery paths. In the first phase, for any give OD (i.e., Origin-Destination) pairs, we aim to identify the shortest delivery paths and also with the corresponding travel times by mining the historical taxi trajectory data offline. In the second phase, using the obtained paths and travel times as the reference to guide the adaptive path-finding, we propose an online taxi scheduling algorithm that aims to discover the near-optimal path iteratively upon the newly incoming taxi ride requests. Finally, with the large-scale taxi trajectory data collected from real life and the package delivery requests generated artificially, we conduct extensive experiments to verify the performance of CrowdDeliver. The experimental results are promising and show that more than 85% packages can be sent to their destinations within 8 h, with an average taxi relay of 4.2.

Part of this chapter is based on a previous work: C. Chen et al., “Crowddeliver: Planning City-Wide Package Delivery Paths Leveraging the Crowd of Taxis,” in IEEE Transactions on Intelligent Transportation Systems, vol. 18, no. 6, pp. 1478–1496, June 2017, doi: https://doi.org/10.1109/TITS.2016.2607458.