Multi-objective Crowd-sensing Task Allocation Based On Satin Bowerbird Optimizer

Compared with the traditional Internet of Things,the Crowd-Sensing Network has the advantages of low cost and wide coverage,so it has attracted the attention of scholars and has been widely used.Task allocation is a key problem that the Crowd-Sensing Network needs to solve.It is a NP-hard combinatorial optimization problem,and it is difficult to solve it with traditional optimization algorithms.Crowd-Sensing task allocation often needs to optimize the values of multiple targets at the same time.The swarm intelligence optimization algorithm is suitable for solving NP-difficult combinatorial optimization problems because of its low solving cost and strong optimization ability.The Satin Bowerbird Optimizer is a new type of the swarm intelligence optimization algorithm,which has better convergence and higher accuracy when solving single-objective NP-hard problems.Therefore,this paper proposes a multi-objective Satin Bowerbird Optimizer,and then establishes two multi-objective Crowd-Sensing task allocation models,and uses a multi-objective Satin Bowerbird Optimizer to solve these two models.In order to obtain a multi-objective optimization algorithm with good performance,this paper proposes a multi-objective Satin Bowerbird Optimizer based on non-dominated solution sorting and crowding-distance calculation.Non-dominant solution sorting is introduced to reduce the time complexity of solution sorting,and the crowding-distance calculation ensures the diversity of the target population,speeds up the convergence speed of the algorithm,and ensures that the understanding set is evenly distributed.Compare the proposed multi-objective Satin Bowerbird Optimizer with three classic multi-objective optimization algorithms such as the multi-objective genetic algorithm(NSGA-II),the multi-objective covariance matrix evolution algorithm(MO-CMA-ES),the strength pareto evolution algorithm(SPEA2),the multi-objective genetic algorithm based on reference point(NSGA-III)and the multi-objective particle swarm optimization algorithm based on balance fitness(NMPSO)in the standard test functions CF1,CF3,CF4 and IMOP2.A variety of experimental evaluation indicators verify that the multi-objective Satin Bowerbird Optimizer proposed in this paper has better solution performance.In the Crowd-Sensing Network,considering the economic benefits of the task platform and the cost of task participants,for the purpose of maximizing the quality of task coverage of the Crowd-Sensing system platform,while minimizing task incentive costs and task resource loss,establish a multi-objective Crowd-Sensing task allocation model with task coverage and incentive-loss as the optimization goal.Different from the traditional task allocation model,the model in this paper first uses the task evaluation function to determine the priority of the task,initializes the task allocation sequence and then solves the target problems.Experiments show that the multi-objective Satin Bowerbird Optimizer in solving the task allocation model can obtain a better solution to the target problems than the algorithm NSGA-II,MO-CMA-ES,SPEA2 and NSGA-III.Considering that the platform not only needs to cover the tasks as much as possible to every participant,but also requires participants to complete tasks as actively and better as possible while obtaining lower profits,so as to achieve the purpose of improving the conversion rate of economic benefits by the Crowd-Sensing system platform,a multi-objective Crowd-Sensing task allocation system model based on task completion and task profit is established.On the basis of the previous model,the introduction of the concept of task completion can effectively improve the low quality of task completion in the group intelligence perception system platform.Experiments show that the multi-objective Satin Bowerbird Optimizer to solve the problem model can obtain a better objective function solution set.