Research On Task Allocation And Participant Recruitment Strategy In Mobile Crowdsensing

Mobile crowdsensing has become a new type of distributed computing perception mode.The intelligent terminal equipment carried by mobile users is used as a perception carrier,and with its powerful sensing,storage,computing,and communication capabilities,a sensing network is established to collaborate to complete complex large-scale Perceptual tasks.Compared with traditional sensor networks,mobile crowdsensing has obvious advantages and is widely used.Crowdsensing will also face some challenges such as participant selection,incentive mechanism,task allocation,privacy protection,etc.This article mainly studies the task allocation and participant recruitment problems in two different application scenarios,and models the problems respectively.In the offline/online scenario,a task allocation and participant recruitment algorithm is proposed.Finally,the performance of the new algorithm is simulated and verified based on real and simulated data sets.The main work and innovation of this paper are as follows:(1)Based on the task assignment problem in mobile social network scenarios,a multi-copy cooperative assignment algorithm based on task priority is proposed.In this scenario,task allocation and data return are realized through social relationships and encounter rules between users.In this method,the user's reputation ability evaluation and reverse incentive based on user contribution value are combined with the user's historical reputation ability and contribution value to evaluate the meeting user in real time in task allocation process.The participants with high contribution value and reputation ability are preferentially selected,and the reputation ability value will be updated with the returned results after completing the task.Task levels are assigned to users with the lowest pre-allocation index in order of priority.Among them,multiple copies of high-priority tasks will be completed multiple times,and the results will be evaluated for data quality.The final result will be selected by majority vote.The corresponding allocation algorithm is proposed in offline/online scenarios and verified by simulation.Experiments show that the proposed algorithm reduces the average task completion time and improves task completion efficiency under the premise of data quality assurance.(2)Based on the problem of participant recruitment in large-scale user perceived scenarios,a dynamic participant recruitment algorithm based on spatiotemporal multitasking is proposed.In this scenario,there are a large number of mobile user pools.Sensing multitasking is heterogeneous and reaches dynamically in different time and space.The algorithm proposed in this paper proposes user and task coverage evaluation methods.Recruitment strategies are designed in both offline and online scenarios.The offline recruitment algorithm assumes known task distribution location information and meeting users.According to the user's historical coverage and known task coverage,Rate to calculate overall user coverage.In the online recruitment algorithm,the task arrival time,completion status and meeting user status are unknown.Based on the user arrival time at the current moment,the start and end of the task,and the remaining time,the task coverage and overall user coverage are dynamically calculated,pre-allocate participant groups online in advance by buffering data budgets that have not reached the coverage of tasks.It is preferred to select the user with the largest overall coverage as a participant in offline/online algorithm.Finally,it is verified by experiments that the algorithm in this paper can minimize the number of participants and the cost of the platform,maximize the level of task coverage,and improve the efficiency of task completion.