Research And Application Of LTE Network Power Optimization Technology

TD-LTE network optimization aims to ensure the quality of service in actual network by allocating the network's resources reasonably. Therefore, analyzing the actual network's data, optimizing various parameters in LTE network and allocating wireless resources rationally are important parts of network optimization.In the TD-LTE network, the transmit power of the cell is the final wireless network's resource, so the power control is the most fundamental and the most important way of network optimization. By rationally allocating the transmit power of cells to improve the network's situation of coverage and handover, to reduce the interference and to improve quality of service in network is of great significance.The optimization of TD-LTE cell's transmission power needs to consider several optimization indexes at the same time, and also needs to satisfy multiple constraints.The problem of power optimization in LTE network is a typical multi-objective and multi-constrained optimized problem. Multi-objective evolutionary algorithms are more suitable to solve those problems. Therefore, this paper applies MOEA/D algorithm to solve the problem of power optimization in LTE network.In this paper, firstly to analyze the multiple objectives and constraints of power optimization in LTE network, and then to analyze the coverage of LTE network with the characteristics of actual data in network. Then, to determine the six sub-targets which contains the minimization of sum of cell's power, the minimization of interference, the minimum number of over coverage, the minimum number of weak coverage, the minimum number of overlapping coverage, and the minimum number of handoffs. The model of TD-LTE network power optimization is established, and the process of evolutionary algorithm based on decomposition is clarified. In this paper, the individual genes are encoded as one-dimensional integer vector, which each vector represents a power-distribution scheme and which is individual that is genetically evolved. The multi - objective problem is decomposed into six single -objective problems, and each sub - objective is obtained by weighted combination of the original multi - objective. In the MOEA/D framework, each sub-problem corresponds to a sub-population,and sub-populations will evolve independently,and the evolution contains selection, crossover, mutation and elitist retention. However, in order to prevent the algorithm from falling into local optima,the size of the evolutionary sub-population's neighbors, the size of the crossover probability, and the size of the mutation probability are adjusted Dynamically that are helpful to the convergence rate of the algorithm. In addition, the local search of small-scale network is implemented for small data sets, and the optimal power allocation scheme for small data networks is proposed. Finally, in order to select the optimal power allocation scheme according to the sub-goal preference, the optimal strategy of Pareto and combinatorial algorithm based on sub-goal preference is achieved.Based on the pivotal technologies above, the software system of TD-LTE network power optimization is designed and implemented. The software is based on the C #language and relies on the .NET platform, uses SQL Server database and XML documents for data storage. Finally,the software system is tested and analyzed comprehensively with the actual network data in the laboratory. Through the test,the power allocation scheme can be obtained under various conditions. The test results verify the rationality and effectiveness of the algorithm. Testing the backstage management function is to verify the practicality of the platform. Through the above work and then the software system can normally put into actual use of the existing network.