| The application of 4G mobile communication technology provides great convenience and rich content for daily life.With the growth of mobile entertainment application market,the demand for mobile data communication among consumers is expanding day by day.In order to meet the demand of QoS in mobile communication service,telecommunication network optimization becomes more important.Telecommunication station planning is an important step in the optimal project of mobile communication network,which affects the quality of communication service in the whole network cell area.Generally,engineers use traditional database for manual planning,because the large amount of data is easy to lead to low efficiency.Therefore,the optimization and improvement of mobile station planning method has practical significance and research value.In recent years,relevant field has been studying and applying big data extensively.Spark,as a popular big data processing platform,with features of RDD,calculation based on memory,good compatibility with Yarn.With libraries such as MLlib and Streaming,making it extensible in machine learning,real-time flow data calculation.This thesis firstly discusses the related technology and development status of mobile communication station planning,and then discusses Spark's development and structure,including RDD,DataFrame,DataSet,Yarn,the workflow of Yarn-Cluster mode,and MLlib library.The next section is the introduction of spatial clustering algorithms.Then design the overall architecture of the station planning system,system is divided into the collection layer,analysis layer,processing layer and application layer from bottom to top,explain the content of each layer,the interface and communication between the layers.In the following part,introducing the main functions including mobile communication data module,site planning module and big data module.Finally,using the data of GuangZhou area to do experiments on system and show effects of station planning. |
PDF Full Text Request