Research On Participatory Multi-task Allocation Mechanism In Mobile Crowdsensing Network

With the continuous expansion of the application field of mobile crowdsensing network(MCS),more and more task publishers use mobile crowdsensing network to collect sensing data,and the multi-task allocation scenario is becoming more and more common.In this paper,we mainly consider the participatory sensing,which takes the task publisher as the center and the sensing time as the boundary,according to whether the sensing task has special requirements for the sensing time(For example,the maximum response time of the task),we divide the task allocation scenarios of MCS into two categories: time sensitive and time insensitive multi-task allocation scenarios.For these two scenarios,the main contributions and innovations of this paper are as follows:(1)For the time sensitive multi-task assignment scenario,task publishers hope to get as high information quality as possible through lower sensing cost and time cost.However,the existing task allocation research rarely considers the three optimization objectives at the same time a In view of this situation,this paper presents a Particle Swarm Optimization Algorithm based on Iteration Number and Fitness Value(INFV-PSO).On this basis,the task allocation mechanism is constructed,which achieves a balance between the essential sensing cost,sensing time and information quality,and then allocates the best user subset for each task.At the same time,the inherent premature problem of standard particle swarm optimization is solved by using dynamic inertia weight based on index and learning factor based on adaptive adjustment of fuzzy rules.Simulation results show that,compared with the existing algorithms,the proposed algorithm reduces the average budget by 11.86%?12.65%,the average response time by 8.6%?12.67% and the task completion rate by 4.1%?8.52% without reducing the average information quality satisfaction.(2)For the time insensitive multi-task assignment scenario,task publishers hope to get as high information quality as possible through lower sensing cost.Most of the existing studies use greedy algorithm,which is easy to fall into the local optimal solution,and rarely consider the existence of multiple task publishers.In view of these situation,this paper presents a Genetic Algorithm based on Greedy Algorithm Initialization for N Task Publishers(N-GAI-GA).On this basis,the task allocation mechanism is constructed,the two-objective optimization problem of sensing cost and information quality is transformed into single objective optimization problem which maximizes the quality reward ratio by using the strategy of mutual relationship.N-GAI-GA uses the global optimization search performance of genetic algorithm to solve the local optimal solution problem of greedy algorithm.At the same time,for the problem of resource shortage caused by N task publishers,a heuristic crossover method is proposed to adapt to the one to many relationship between users and tasks.Simulation results show that the total quality reward ratio(TQR)of the proposed algorithm is improved by 9.82%?11.53%.