Low-Complexity Recruitment for Collaborative Mobile Crowdsourcing Using Graph Neural Networks

Collaborative mobile crowdsourcing (CMCS) allows entities, e.g., local authorities or individuals, to hire a team of workers from the crowd connected people, execute complex tasks. In this article, we investigate two different CMCS recruitment strategies allowing task requesters form teams socially and skilled workers: 1) platform-based strategy where platform exploits its own knowledge about 2) leader-based designates group leader that recruits suitable given social network (SN) neighbors. We first formulate as an integer linear program (ILP) optimally forms according four fuzzy-logic-based criteria: level expertise; relationship strength; 3) cost; 4) recruiter’s confidence level. To cope with NP-hardness, design novel low-complexity approach relying on graph neural networks (GNNs), specifically embedding clustering techniques, shrink workers’ search space afterwards, exploiting metaheuristic genetic algorithm select appropriate workers. Simulation results applied real-world data set illustrate performance both proposed approaches. It is shown our GNN-based achieves close performances those baseline ILP significant computational time saving ability operate large-scale platforms. also compared strategy, more but lower SN relationships higher cost.