Design And Implementation For PCI Network Planning Of LTE Network Based On Genetic Algorithm

LTE(Long Term Evolution)network,use the PCI(Physical layer Cell Identity)to identify different cells,so PCI for LTE cells has a very important role,a total of 504 PCI code,PCI planning if unreasonable,the user equipment may camp on the wrong cell,will bring serious interference problems,and the current increase in base station density and network size of the surge in demand for PCI planning urgent.The theoretical model of LTE network PCI planning is how to allocate the limited PCI to the cell in the network under the certain constraint c ondition,so that the whole interference of the network is minimized.Some existing planning methods,due to planning data source and planning algorithm reason,The existence of the applicable network size is small,the planning error and other issues.Based on this background,this paper studies the PCI network planning model of LTE network,compares the adaptability of different algorithms to these problems,and adopts genetic algorithm as the global planning algorithm of this project,and designs an improved genetic algorithm for PCI planning.The algorithm of the PCI overall interference value of the network as a fitness function,and in order to avoid the common binary coding in the variation of the defects,the encoding method using Gray code,in the algorithm of population selection phase by adding simulated annealing idea,using improved Roulette selection method,dynamic adjustment of selection pressure,cross-use of single point crossover,mutation method using multiple random selective mutation method,planning data source based on inter-cell correlation measurement data structure of two-dimensional interference matrix.In the system implementation,the GA algorithm is encapsulated,and by means of GIS platform and database,the PCI planning syste m is developed to realize the automatic planning of PCI.And in the actual mobile communication LTE network planning experiments were verified.This method combines the communication knowledge with the computer algorithm,and adopts the improved Gray code coding method in the genetic algorithm,the simulated annealing design by the roulette wheel selection method,and the two-dimensional interference matrix in the cell correlation,which can realize the automatic unattended Automated PCI planning.The experimental results show that the performance of the system is fully meet the design requirements.In terms of performance,the system has the characteristics of high stability,fast convergence speed and strong global search ability.The PCI network bad counting,drvie test index and performance index are improved obviously.Among them,the interference value is improved by 48.47% and the downlink throughput of PDCP is higher than 20 M / s.The performance of CQI SINR and dropped call rate is improved obviously.The experiment has achieved the desired effect.