Acceptance-aware multi-platform cooperative matching in spatial crowdsourcing

With the development of sharing economy, multi-platform cooperative matching (MPCM) is becoming popular as it provides an effective way to cope with the supply-demand imbalance in spatial crowdsourcing (SC). While cooperation between two SC platforms in MPCM has been intensively studied, competition among multiple SC platforms is largely overlooked by existing work. In particular, an idle worker may be requested by multiple platforms simultaneously, but he/she can only accept some of them due to capacity constraints. This partial acceptance will decrease the revenue of some platforms and thus should be addressed properly. Towards this goal, we investigate in this paper the problem of acceptance-aware multi-platform cooperative matching. We first design an algorithm called BaseMPCM to predict the acceptance rate of workers and calculate the utility scores of task-and-worker pairs. Considering that in BaseMPCM, the platforms make the decision from their own benefits, and this may lead to a sub-optimal total revenue, we further design an algorithm called DeepMPCM to predict the action of other platforms and calculate the utility scores globally. Extensive experiments on real and synthetic datasets demonstrate the effectiveness of our algorithms.