Query Optimization And Analysis Of Trajectory Data Based On Distributed Computing Platform

The rapid development of mobile Internet and the wide application of mobile terminals have given rise to a large number of trajectory data.These data record spatio-temporal mobility of mobile objects and contain the behavior information and interaction information of the individual or group.It has important value for urban planning,commercial location selection,traffic congestion detection,etc.However,how to quickly retrieve trajectory data and find valuable information is a challenging problem when facing with a large amount of trajectory data.Therefore,this paper makes in-depth study,and its contributions are as follows:(1)The data locality of the partitioning technology reduces the range query efficiency of the trajectory data.Therefore,this paper proposes a partitioning technique that can improve the range query efficiency of the trajectory data.The partitioning technology also has the following advantages: the trajectory data is directly partitioned without preprocessing;when the new trajectory data arrives,the trajectory data that has been partitioned does not need to be repartitioned,so it can be better adapted to the online task.Finally,it is verified by experiments that the range query based on this partitioning technology has higher query efficiency than other partitioning technologies.(2)In view of the important reference value of the trajectory data of mobile objects for commercial location selection,this paper proposes an optimal location search model based on trajectory data.The model considers the difference between the sampling points of different users and the interaction between the sampling points of the same user so as to adapt to different application scenarios.In order to further improve the computing performance,this paper proposes two distributed deployment schemes for this model.Finally,the high efficiency of the distributed deployment scheme is verified by experiments,and the application scenarios of the two schemes are analyzed and summarized based on the experimental results.(3)Aiming at the important value of trajectory data analysis for real-time traffic congestion detection,this paper proposes two methods for real-time detection of traffic congestion based on trajectory data: one is distributed DBSCAN density clustering;the other is distributed topological analysis of trajectory data.Finally,three real trajectory data sets are used to simulate the real-time traffic congestion detection process of the two methods on the streaming batch engine,and their application scenarios are analyzed and summarized according to the experimental results...