Three-dimensional Stable Task Assignment in Semi-opportunistic Mobile Crowdsensing

In semi-opportunistic mobile crowdsensing (SO-MCS), workers are asked to provide the matching platform with multiple paths they find acceptable between their starting locations and destinations in order to alleviate the problem of poor coverage in opportunistic MCS without forcing them to take potentially much costly and hence undesirable paths as in participatory MCS. While these alternative paths open up new assignment possibilities between workers and tasks, they also make it more challenging to find a stable or preference-aware task assignment (TA), as they bring a new dimension to the TA problem (i.e., workers/paths/tasks instead of workers/tasks as in previous work), and introduce complex requirements to achieve stability by satisfying user preferences. In this paper, we formally define the stability conditions for three-dimensional task assignments in SO-MCS, and propose two polynomial-time TA algorithms: an exact algorithm for SO-MCS systems with uniform worker qualities, and a c-approximate algorithm for general SO-MCS systems, where c is the number of the acceptable paths of the worker with the largest set of acceptable paths. Through extensive simulations, we demonstrate that the proposed algorithms significantly outperform the state-of-the-art TA algorithms in terms of stability (or user happiness) in most scenarios.