A Cross-Region-based Framework for Supporting Car-Sharing

Abstract

With the rapid development of mobile Internet and sharing economy, carsharing has attracted a lot of attention around the globe. Many popular taxi-calling service platforms, such as DiDi and Uber, have provided carsharing service to the passengers. Such carpooling service reduces the energy consumption while meeting passengers’ convenience and economic benefits. Although numbers of algorithms have been proposed to support carsharing, the computing efficiency and matching quality of these existing algorithms are all sensitive to the distribution of passengers. In many cases, they cannot effectively and efficiently support carsharing in an on-line way. Motivated from the aforementioned issues and challenges, in this paper, we propose a novel framework, namely, Cross-Region-based Task Matching (CRTM) for supporting carsharing for smart city. Compared with existing algorithms, CRTM analyzes and monitors regions having multitudes of tasks for car sharing among users. In order to achieve this goal, we first propose a new machine learning-based algorithm to find a group of regions which contain many tasks. Then, we propose a novel index, namely, Included Angle Partition-based B-tree (IAPB), for maintaining tasks such as (i)whose pick-up points are contained in these regions, (ii) that may pass this kind of regions. Thirdly, we propose three buffer-based matching algorithms for cross-region-based task matching. Experiment results demonstrate the significant superior performance of the proposed algorithms in terms of energy saving and overall cost minimization.