QQ客服 官方微博 反馈留言

	Achieving Stable and Optimal Passenger-Driver Matching in Ride-Sharing System
	Yixuan Zhong; Lin Gao; Tong Wang; Shimin Gong; Baitao Zou; Deliang Yu
	2018
会议日期	2018
会议地点	成都
英文摘要	Ride-sharing systems enable individual car owners with idle time to provide commercial taxi-like services via an online platform. By crowdsourcing a large population of individual car owners, it can provide more flexible services with a lower serving cost, comparing with the traditional taxi system. Due to the autonomous nature of car owners (drivers), a decentralized driver dispatching algorithm that can achieve a stable (self-motivated) and optimal passenger-driver matching is highly desired for a ride-sharing system. In this paper, we will study such a driver dispatching algorithm systematically. We first show that the optimal passenger-driver matching achieved by the centralized driver dispatching algorithm is often not stable, in the sense that some drivers and passengers may break with their matched partners and form new matching pairs. To this end, we introduce a virtual order fee on each passenger (which the platform will charge the drives who want to serve the passenger) to motivate the behaviors of drivers. Specifically, we propose a novel auction-based decentralized driver dispatching algorithm, where each driver proposes the most profitable passenger that he wants to serve, by jointly considering the potential profit that he can achieve and the order fee that he needs to pay from/to serving each passenger. The virtual order fee on a passenger will be gradually increased when multiple drivers want to serve the passenger, until there exists only one driver who is willing to serve. We analytically show that such a decentralized driver dispatching algorithm will converge to an equilibrium (stable) outcome, which achieves the optimal passenger-driver matching (i.e., that maximizes the social income of the whole system). Simulation results further show how the converging speed and the achieved social income change with the system parameters such as the step size of order fee increasement. Moreover, it is easy to implement the proposed distributed algorithm in a practical system.
语种	英语
内容类型	会议论文
源URL	[http://ir.siat.ac.cn:8080/handle/172644/14141]
专题	深圳先进技术研究院_数字所
推荐引用方式 GB/T 7714	Yixuan Zhong,Lin Gao,Tong Wang,et al. Achieving Stable and Optimal Passenger-Driver Matching in Ride-Sharing System[C]. 见:. 成都. 2018.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。