Correlated Friends' Impacts in Social-crowdsensing

Social sensing is a typical application of the crowdsourcing system. With the consideration of system timeliness, flexibility, and stability, it could not be more natural to build a self-organized, distributed, and cross-platform crowdsourcing system. Social-crowdsensing (SC) is the first attempt. In SC, a huge sensing task is gradually partitioned into smaller pieces, and the pieces are propagated to potential workers via stochastic social contacts. During these contacts, allocating the workload is a critical problem, which affects the work's completion time and system resource utilization. By analyzing real data, we notice that the times of social contact occurrences are partially correlated. Whether it is necessary to purposely incorporate workers' correlation into the decision-making phase of workload allocation becomes an open question. In this paper, we systematically study the impacts of users' correlated behaviors.