Max-Min Fairness Multi-task Allocation in Mobile Crowdsensing

Abstract

Mobile Crowdsensing (MCS) has become a new paradigm of collecting and merging a large number of sensory data by using rich sensor-equipped mobile terminals. Existing studies focusing on multi-task allocation with the objective of maximizing the social utility may result in the problem of unbalanced allocation due to the limited resources of workers, which may damage the social fairness, and requesters who suffer unfairness will choose to leave the system, thereby destroying the long-term stability of the system. To address this issue, we introduce max-min fairness into the design of a novel fairness-aware incentive mechanism for MCS. We first formalize the max-min fairness-aware multi-task allocation problem by using the sensing time threshold of tasks as a constraint. By modeling the max-min fairness-aware multi-task allocation problem as a Stackelberg game consisting of multi-leader and multi-follower, we next compute the unique Stackelberg equilibrium at which the utilities of both requesters and workers are maximized. Then, we design a greedy algorithm to achieve max-min fairness while meeting the sensing time threshold required by the task. Finally, simulation results further demonstrate the impact of intrinsic parameters on social utility and price of fairness, as well as the feasibility and effectiveness of our proposed max-min fairness-aware incentive mechanism.