Load Balancing In Heterogeneous Networks Based On Data Mining Algorithms

With the popularity of mobile Internet and the concept of'Internet +'and the widespread of smart devices,the mobile traffic will meet a dramatic growth.Furthermore,with the transformation of people's living habits,more and more traffic occurs on hot spots,which causes network load imbalance.In order to achieve load balance and improve network quality of service,this paper effectively analyzes the temporal and spetial dynamics of network traffic by utilizing data analysis and data mining tools,proposes the allocation of heterogeneous networks resources and utilizes the techniques in actual network.These issues are urgent to tackle.This thesis is focused on two key problems,i.e.,the dynamic allocation of heterogeneous network resources,improving the quality of communication service of the network.This paper proposes the dynamic traffic offloading scheme and the actual network configuration solutions by exploring the historical data coolected by a network operation.The research consists of three parts.The first part is analyzing the dynamic characteristics of traffic by utilizing data datamining and data analysis tools.Further,fitting out the arrival model of the data packets based on the network operation data.The second part initially proposes the concept of traffic offloading region based on the queuing theory that acer mode.On this basis,a dynamic traffic offloading scheme is proposed to guide network resource allocation.The last part applys cell range extension technology to make sure the specific amount of traffic to be offloaded to small cells.In order to improve the communication quality of service of edge users,interference coordination is applied to reduce the interference from macro base stations.Finally,the research is concluded by pointing out the deficiencies of the the proposed scheme and indicating the direction for improvement in future work.