Research On Network Self-optimization Technologies Based On SON In 5G Communication System

With the development of wireless communication technology, the consistent growth of customer services, and the substantial quantity of intelligent terminal systems, the existing network is confronted with huge challenges. In order to keep pace with the rapid development of mobile services in the future, and to adapt to the rapid growth of mobile communication data traffic in a better way, the fifth generation communication system, also known as 5G network, emerged. For the traditional mobile communication network, the operation and maintaining of the network usually rely on human works, which result in the waste of labor and cost. In addition, those optimizations, done by human, fail to meet the standard. In the future, 5G network will be faced with many challenges in terms of deploy, operation and maintaining because of the various wireless access technologies. Those various network nodes have different coverage abilities and complex inter-relationships, which pose a lot of challenges. Therefore, the intelligent network remains a key technology in 5G network. Based on 5G communication system, the theory and key technology of self-organizing network are discussed in this thesis, focusing on the detailed comparison of two aspects—load balancing and base station dormancy-to further demonstrate the self-optimization technology of self-organizing network. The works are as following:Firstly, a load balancing algorithm based on network spectrum efficiency optimization is proposed. If the load balance is broken, the availability of system resources will be reduced. However, if the load balance is the only target, some user terminals will end up connecting to the poor access points, resulting in bad user experience and reducing the system spectrum utilization. In this way, two issues are considered, the load balance and the improvement of spectrum efficiency, as a multi-objective optimization problem. Then through ratio method, this problem can be simplified as one single-objective optimization problem,and then solved by quantum particle swarm optimization algorithm. The simulation shows that the algorithm mentioned above can improve the spectrum efficiency and maintain the load balance at the same time.Secondly, the dormancy mechanism of base station based on a non-cooperative game theory is put forward. It is well acknowledged in the industry that ultra-dense network is one of the key technologies in 5G network. With the substantial deployment of low power access points, the overall network energy efficiency becomes a vital indicator in measuring the long-term economic benefits of 5G communication system. Base station dormancy remains an effective way to reduce the overall energy consumption of the network by closing the light-loaded access point, and diverging users to other neighboring access points. Therefore, this thesis proposes the dormancy mechanism based on non-cooperative game theory, introducing the first score auction model. In this model, the quasi dormancy base station and its nearby base stations are modeled as the auction model, and all user terminals in the quasi dormant base station act as goods in an auction, switching to the target base station through auction, and then the quasi dormant base station can lie dormant. The simulation shows that the dormancy mechanism of base station based on the non-cooperative game theory, proposed in this thesis, can reduce the energy consumption of the network and improve its energy efficiency.