Optimizing Of Power Configuration In TD-LTE Networks

TD-LTE networks guarantee the network service quality and user experience of the system through optimization of wireless network resources.It is an important part in TD-LTE networks optimization to continuously adjust and optimize various parameters in TD-LTE networks and allocate the wireless network resources reasonably by analyzing actual network data.The cell RS(Reference Signal)transmitting power is a type of important wireless network resource in TD-LTE networks.Controlling cell RS transmitting powers is a very essential method of network optimization.In large-scale practical TD-LTE networks,the reasonable configuration and optimization of cell RS transmitting power can effectively handle TD-LTE network coverage and intra-cell interference,and it is of great significance to improve the network service quality and enhance the user experience.Optimization of cell RS transmitting power aims to optimize multiple objectives simultaneously under the premise of satisfying certain constraints,which can be reduced to a typical multi-objective and multi-constraint optimization problem.One effective method to solve this kind of problem is the multi-objective evolutionary algorithm.In this thesis,we use the MOEA/D(Multi-Objective Evolutionary Algorithm based on Decomposition)to solve multi-objective optimization of cell RS transmitting power in TD-LTE networks,and introduce a cell-type-oriented optimization strategy and a parallel evolutionary algorithm into MOEA/D.The MOEA/D-based power optimization scheme is follows.First,the cells to be optimized are classified as the ATU optimized cells and MRO optimized cells.Different optimization objectives are determined for the different types of optimized cells,and a mathematic model of constrained multi-objective optimal assignment of transmitting power is established.Then,the cell MRO weak coverage is analyzed to determine the lower limit and adjusting direction of RS transmission power of the cell to be optimized.Meanwhile,the cell KPI and handover data are used to determine the cell load level and load balancing cell pairs.Then,one-dimensional vector floating-point coding is used to represent the individual genes of the cell RS transmission power configuration scheme.The weighted sum method is taken to decompose the multi-objective optimization of RS transmission power assignment problem into multiple single-objective optimization problems.The uniform design method is employed to design the weight matrix of sub-problems.The optimization goal of each sub-problem is the weighted sum of different sub-objective after normalization on different weight vectors.Each sub-problem constructs a subpopulation and selects the neighbor populations for its sub-population.Aiming at the optimization of RS transmission power assignment,the special selection,crossover and mutation evolution operators are designed.The adjacent populations provide information for the evolution of the sub-populations.The sub-problem dynamically adjusts the sizes of the neighboring populations,and independently executes the evolution operators,so that the sub-problems can be evolved from generation to generation.And the evolution process in MOEA/D is parallelized to speed up the evolutionary calculation in the problem solving process.The result of MOEA/D-based optimization are represented as a set of non-dominated solutions,each of which denotes a particular RS power allocation.After evolution,all the non-dominated solutions are further optimized by the scheme of backtracking-based local search,so as to improve the quality of the non-dominated solutions and make its more suitable for actual network applications.Finally,one or more solutions from the non-dominated solution set are selected,by using the intersection method and the comprehensive scoring mechanism,as the final optimized assignment of RS transmission power.Based on the aforementioned technologies,a software system for power optimization and configuration in TD-LTE networks was developed by using C#programming language under.NET4.5 platform and SQL Server 2012 database.In cooperation with our third-party partners,we obtained the real network data from several domestic running networks in China,and applied the software system to optimize the cell RS transmission power configuration in these networks.By analyzing the convergence of each sub-objective during the evolution process of the algorithm and the improvement of network evaluation indicators,it is verified that multi-objective power optimization technology based on MOEA/D can get better RS transmission power assignment than that in the existing networks in a short time.The aforementioned research work proves that the approach of TD-LTE multi-objective optimization of power assignment described in this thesis can meet the optimization objectives of the real RS transmission power optimization in TD-LTE networks and provide a high-quality RS transmission power configuration scheme,and is a promising way to optimize RS transmission power assignments in practical LTE networks.