A Constrained Many-Objective Mobile Crowdsensing Task Allocation Method Considering Latent Workers

The concept of mobile crowd sensing (MCS) represents a powerful model within the Internet of Things. It perceives different phenomena in the urban environment by taking advantage of the mobility of individuals and the sensors embedded in smartphones. Existing MCS task allocation systems typically assume that workers are in an idle state without any current tasks. However, this overlooks workers who are already engaged in other tasks but are closer to the task location, leading to potential resource wastage. In this article, we refer to such workers as “the latent workers.” To address this issue, we propose a novel model called the constrained many-objective MCS task allocation model considering latent workers (CMaO-MCSTA-LW). This model aims to find a set of compromise solutions under multiple uncontrollable objective environments. To effectively solve this model, we developed a new algorithm called the dual-population constrained many-objective evolutionary algorithm based on the easing strategy and the elite parent strategy (dCMaOEA-E-P), which builds upon the original dCMaOEA-RAE. Experimental results demonstrate that, compared to several other representative algorithms, the proposed algorithm shows superior performance across multiple perspectives, significantly improving task allocation efficiency.