Fairness Counts: Simple Task Allocation Scheme for Balanced Crowdsourcing Networks

Xiaochen Fan,Panlong Yang,Qingyu Li

DOI: https://doi.org/10.1109/msn.2015.32

2015-01-01

Abstract:With the increasing development of mobile networking technologies, optimization methods for efficient task assignment plays a key role for mobile crowdsourcing process. However, what hiding behind the strategies are solutions to motivate users for participation, which reveals a fundamental problem: the fairness issue of crowdsourcing system. Since the participators are human beings with intensive interest for obtaining benefits, it is reasonable to build a sustainable crowd with guaranteed fairness among users. Thus in this study, we investigate the fairness issue in mobile social network, which could be more complicated when uncontrollable mobile users are concerned. The intuitive solution is, if the tasks could be effectively assigned among users in a balanced way, the fairness could be guaranteed. Unfortunately, there is still a big challenge for this issue, because it's difficult to acquire accurate global information of task loading, which is highly dynamic and distributed. By leveraging the power of two random choices, which is based on the balls and bins theory, we develop a lightweight scheme to allocate tasks. Indeed, we proposed a heuristic algorithm to achieve balanced task allocation effectively with O(1) complexity. To the best of our knowledge, it is the first effort for incorporating fair load balancing in pure distributed mobile crowdsourcing systems. Our extensive evaluation results validate our task offloading algorithm, showing that the proposed scheme outperforms the random choice method.