Crowdsourcing Sensing to Smartphones: A Randomized Auction Approach

被引:36
作者
Li, Juan [1 ,2 ]
Zhu, Yanmin [1 ,2 ]
Hua, Yiqun [1 ,2 ]
Yu, Jiadi [1 ,2 ]
机构
[1] Shanghai Jiao Tong Univ, Dept Comp Sci & Engn, Shanghai 200240, Peoples R China
[2] Shanghai Key Lab Scalable Comp & Syst, Shanghai 200240, Peoples R China
关键词
Mobile crowdsourcing; incentive mechanism; randomized auction; diversity; quality; truthfulness; rational; social cost; SYSTEMS; TASKS;
D O I
10.1109/TMC.2017.2653774
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Crowdsourcing to mobile users has emerged as a compelling paradigm for collecting sensing data over a vast area for various monitoring applications. It is of paramount importance for such crowdsourcing paradigm to provide effective incentive mechanisms. State-of-the-art auction mechanisms for crowdsourcing to mobile users are typically deterministic in the sense that for a given sensing job from a crowdsourcer, only a small set of smartphones are selected to perform sensing tasks and the rest are not selected. One apparent disadvantage of such deterministic auction mechanisms is that the diversity with respect to the sensing job is reduced. As a consequence, the quality of the collected sensing data is also decreased. This is due to failure to exploit the intrinsic advantage of the large set of diverse mobile users in a mobile crowdsourcing network. In this paper, we propose a randomized combinatorial auction mechanism for the social cost minimization problem, which is proven to be NP-hard. We design an approximate task allocation algorithm that is near optimal with polynomial-time complexity and use it as a building block to construct the whole randomized auction mechanism. Compared with deterministic auction mechanisms, the proposed randomized auction mechanism increases the diversity in contributing users for a given sensing job. We carry out both solid theoretical analysis and extensive numerical studies and show that our randomized auction mechanism achieves approximate truthfulness, individual rationality, and high computational efficiency.
引用
收藏
页码:2764 / 2777
页数:14
相关论文
共 43 条
[1]  
[Anonymous], 2001, Approximation algorithms
[2]  
[Anonymous], 2009, CONVEX OPTIMIZATION
[3]  
[Anonymous], 1972, P COMPLEXITY COMPUTE
[4]  
Birrell E., 2011, P IJCAI, P67
[5]   Randomized metarounding [J].
Carr, R ;
Vempala, S .
RANDOM STRUCTURES & ALGORITHMS, 2002, 20 (03) :343-352
[6]  
Carr R, 2000, PROCEEDINGS OF THE ELEVENTH ANNUAL ACM-SIAM SYMPOSIUM ON DISCRETE ALGORITHMS, P116
[7]  
Chen C., 2014, Proc. Second AAAI Conf. Hum. Comput. Crowdsourcing (HCOMP 2014) TRACCS, P30
[8]   BikeNet: A Mobile Sensing System for Cyclist Experience Mapping [J].
Eisenman, Shane B. ;
Miluzzo, Emiliano ;
Lane, Nicholas D. ;
Peterson, Ronald A. ;
Ahn, Gahng-Seop ;
Campbell, Andrew T. .
ACM TRANSACTIONS ON SENSOR NETWORKS, 2009, 6 (01)
[9]  
Feng ZN, 2014, IEEE INFOCOM SER, P1231, DOI 10.1109/INFOCOM.2014.6848055
[10]  
Gao L., 2015, P IEEE INFOCOM, P1803