Research On Mobile Crowdsourcing Task Assignment Method Based On User Context Trajectory Prediction

In recent years,with the rapid development of computer and communication technology and the maturity of 5G,the rapid popularization of intelligent devices has been promoted.Various mobile crowdsourcing platforms,such as drip-drip taxis,beauty troupe takeout,Foursquare and so on,have been entered the public's field of vision,covering a wide range of applications.Compared with AMT,Wikipedia and other early traditional crowdsourcing platforms,the task characteristics and worker behavior of mobile crowdsourcing applications are more complex and dynamic.For the task that the task requester issues on the crowdsourcing platform,the "free" and"voluntary" choice of the crowdsourcing worker to accept or reject the task is a basic feature of the crowdsourcing application scenario.Therefore,an effective task allocation strategy is the key factor to determine whether the task can be successfully accepted and completed by the required number of workers.In mobile crowdsourcing applications,there are complex and changeable temporal and spatial factors,such as timeliness of tasks,dynamic distribution of task locations and uncertainty of workers'movement trajectories,and individual factors such as the diversity of workers' hobbies and hobbies pose challenges to the core issue of mobile crowdsourcing tasks.Aiming at mobile crowdsourcing scenarios and the principle that workers voluntarily accept tasks,this paper constructs a mobile crowdsourcing task assignment model based on user trajectory prediction,which mainly includes mobile user trajectory prediction algorithm and task assignment algorithm.The main work is as follows:1.Proposed a mobile user context-dependent trajectory prediction algorithm.By mining the historical traj ectory data of workers,this paper analyses the context-related movement patterns of workers,and constructs the context-related movement rules of workers.Finally,according to the movement rules,the location region that the worker will arrive is predicted,which provides the basis for the next mobile crowdsourcing task assignment.2.Based on the worker's next arrival region obtained by the trajectory prediction method for mobile users,this paper proposes several algorithms for assigning tasks in the region to the worker who arrives in the region,which are based on the task assignment with the largest number of tasks,in order to maximize the number of tasks assigned as much as possible.Task assignment based on the best quality of workers aims to maximize the accuracy of answers collected from assigned tasks.Task assignment based on cost-optimized,this method not only ensures the quality of the collected answers,but also ensures that the distance cost of workers is not too high.3.This paper selects the spatio-temporal data collected by social networking site Gowalla and the real task assignment data collected by the mobile crowdsourcing experimental platform as experimental data sets,and compares and verifies the proposed mobile user trajectory prediction algorithm and task assignment algorithm respectively.The results show that the proposed algorithm has better accuracy and distribution results.Mobile crowdsourcing task assignment method based on user context trajectory prediction,by predicting the trajectory location of mobile users,mobile crowdsourcing platform can assign spatio-temporal tasks to workers most likely to perform the task,and improve the probability of workers accepting and completing tasks,ultimately improve the success rate of space-time task assignment.Finally,this paper chooses real-world data as experimental data set,and compares and verifies the proposed task assignment model based on mobile user trajectory prediction.The experimental results show that the proposed algorithm has better running performance and efficiency.