Research On Task Assignment Mechanism Based On Community And User Characteristics In Mobile Crowdsensing

With the rapid development of wireless communication technology and intelligent terminals,the concept of Mobile crowdsensing has emerged.Mobile crowdsensing utilizes the multi-sensing capabilities of smart terminals(smartphones,tablet computers,etc.)and combines wireless communication technology to complete the collection and transmission of various sensing data.Compared with traditional sensing modes,mobile crowd-sensing has the advantages of low cost,wide range,large scale and strong ubiquity.It has been widely used in air quality detection,map monitoring and modification,indoor positioning,smart city and environmental protection.Therefore,it has very important research significance and application value.Task assignment is a core link in mobile crowdsensing.As the executor of sensing tasks,users are a major factor affecting the design and performance of task assignment algorithms.At present,researchers at home and abroad mainly focus on the user’s spatiotemporal characteristics,task preference and reliability,etc.However,as a member of the society,users’ social attributes cannot be ignored.Users are often in several relatively stable communities,and users in the same community are more connected to each other and tend to have similar spatiotemporal characteristics and preferences.Therefore,how to make full use of the user’s social attributes,propose a reasonable user community detection algorithm,and construct a user community-based task assignment algorithm has become an important topic worth studying.This paper extracts user spatiotemporal features and social features,and proposes a user community discovery algorithm based on user multidimensional features to improve the accuracy of community discovery in mobile crowdsensing scenarios.On this basis,a platformcommunity-user three-layer task assignment model is established,and a task assignment algorithm is designed,which effectively improves the performance of task assignment.The main contributions include:1)Study the task assignment algorithm based on user characteristics in mobile crowdsensing.An assignment model is built that maximizes system revenue based on attributes such as user geographic location,historical task completion records,etc.It is transformed into a bipartite graph maximum weighted matching problem,and a task assignment algorithm is designed in combination with the Hungarian algorithm.On this basis,a task assignment model under budget constraint is further established and transformed into a cost minimization model under benefit constraint and a benefit maximization model under cost constraint using the optimization method.For the former,a greedy task assignment method based on current system benefit is proposed,and for the latter,a task assignment algorithm based on simulated annealing and an improved task assignment algorithm based on performance greedy are proposed,and the performance of the task assignment algorithm is verified through experiments.2)Design the community detection algorithm in mobile crowdsensing.According to the user’s spatiotemporal trajectory,the user’s spatiotemporal characteristics are analyzed,a user movement model is constructed,and a community detection algorithm based on the user’s spatiotemporal characteristics is proposed in combination with the clustering algorithm.According to user social network information,a reasonable similarity of user nodes is designed,a local community detection algorithm based on seed expansion is proposed,and a community merging adjustment algorithm is designed according to the degree of community integration.The community detection algorithm is constructed from the spatiotemporal and social attributes of users respectively,which fully explores the multidimensional attributes of users.3)Present the task assignment mechanism of mobile crowdsensing based on user characteristics and community.Combined with the spatiotemporal characteristics and social characteristics of users,a community discovery algorithm based on user multidimensional characteristics is firstly proposed,and then a platform-community-user three-layer task assignment model is constructed based on user communities.Based on this design,the firstlevel assignment algorithm from the platform to the community and the re-distribution algorithm within the community can fully mine user attributes and improve the overall performance of the mobile crowd-sensing task assignment system.