Utility-Based Task Assignment for ON-based Mobile Crowdsourcing

In this paper we are interested in opportunistic network-based (ON-based) mobile crowdsourcing (MCS), where a requester (called a server) assigns a set of tasks to a pool of workers, and the workers process the assigned tasks for payoff. The key to success such an ON-based MCS is how to design task assignment strategy. The existing task assignment algorithms are primarily designed to maximize the task completion rate or to minimize the makespan. To the best of our knowledge, there is no work on the task assignment with a time-decaying utility model, where task utility decreases over time. Therefore, in this paper, we first introduce the utility-based task assignment problem with a time-decaying utility model for ON-based MCS. Then, the utility-based task assignment (UTA) algorithm is proposed based on the greedy strategy that dynamically assigns tasks to contacted workers. The performance of the proposed scheme is evaluated using a real mobility trace, called CRAWDAD, and the simulation results demonstrate that the proposed UTA outperforms the one of the existing online task assignment algorithms.