Load Balancing Technology For Dense Heterogeneous Network

With the rapid development of a variety of intelligent terminals and a variety of mobile terminal business, the user demand for traffic has doubled. How to deal with short-term bursting of traffic and solving the 1000 times the user traffic demand in the future is a problem of 5G communication technology must be addressed. Facing the limited capacity of the network and the dense coverage of wireless networks,Dense Heterogeneous Networks have been proposed to deal with thousands of times of growth of data traffic, meeting the network demands of high traffic density. However, the distribution of load under hot-spot high-volume scenarios can be more uneven, which greatly reduces the user experience. In this thesis, load balancing technology is studied in Dense Heterogeneous Network to improve system resource utilization, overall system performance and user satisfaction. The following are the main innovations points and research contents of this thesis.In this thesis, we propose a resource estimation scheme for cells and users for the problem of low resource utilization in the process of load balancing. Before load balancing, we first estimate the optimal resource utilization rate per cell and the optimal resource allocated to each user. In this scheme, the heterogeneous network is modeled according to the Cournot model in game theory. The specific system model can be changed according to the different conditions. Then, the Genetic Algorithm is used to optimize the system utility function of the model,calculating the optimal resource utilization rate per cell and the optimal resources allocated to each user, and ultimately improving the system resource utilization during load balancing.In this thesis, we both consider the system resource utilization and user QoS, based on the proposed resource estimation scheme, a heterogeneous network load balancing technology scheme based on QoS guarantee is proposed. Firstly, based on the resource estimation scheme of the cell and the user, the optimum resource utilization of each cell and the optimal resource allocated to each user are predicted. Then, the bipartite graph in graph theory is constructed to realize the relationship between users and cells. Kuhn-Munkras algorithm is used to realize the optimal matching process of bipartite graphs. Simulation results show that the proposed scheme can improve the system resource utilization and system throughput while reducing the outage and improving the service quality.In this thesis, a heterogeneous network load balancing technology scheme based on multi-service coexistence is proposed considering both the system resource utilization and multi-service coexistence. Firstly, this thesis describes the establishment of multi-service coexisting load balancing model. Secondly, by using the resource forecasting scheme proposed by the second chapter, we can get the optimal resource utilization rate of each service in each cell. The optimal resource allocated to different services is based on the predicted value, and the matching relationship between the multi-service and the cell resource is established by the bipartition graph. Simulation results show that the proposed scheme can improve user satisfaction and system resource utilization, system throughput and reduce outage while achieving load balancing.At the end of this thesis, the thesis summarizes the research contents and prospects of the above research.