MOEA/D-based participant selection method for crowdsensing with social awareness

With the explosive popularity of mobile terminal devices, crowdsensing has become a novel paradigm for thoroughly sensing the environment. A crucial issue in crowdsensing system involves that the selection of appropriate participants from a number of mobile users to guarantee completing the sensing tasks. The traditional participant selection methods only consider the profit of the task publisher, resulting in the loss of potential users. Base on this, a novel multi-objective participant selection model is built, with the purpose of guaranteeing the interests of the mobile users by maximizing the total reward of participants and the profit of task publisher by maximizing the overall sensing quality. To solve this combinatorial optimization problem, an improved multi-objective evolutionary algorithm based on decomposition (MOEA/D) is proposed. It utilizes a balance factor to uniform various scales between two objectives. The optimal solution for each objective obtained by greedy algorithm is incorporated into initial population, with the purpose of avoiding falling into the local optima. In addition, a novel mutation operator is developed to enhance the convergence of solutions. The simulation results show that the improved MOEA/D has a significant better performance than the other algorithms for our problem, and the proposed multi-objective participant selection model more fits for the crowdsensing based on social awareness. (C) 2019 Elsevier B.V. All rights reserved.