An Improved NSGA-Ⅱ Algorithm And Its Application Research In 5G Network Planning

The development of multi-objective optimization theory and optimization algorithms has led to the vigorous development of many disciplines,and various combinatorial optimization problems have been characterized by large scale and complexity.Among them,heuristic algorithms have been widely used in many fields such as industrial design and production due to their simple,efficient and convenient parallel processing advantages,and have generated huge economic and social benefits.In particular,heuristic algorithms play an important role in 5G network planning applications.Since the distribution planning of base stations requires comprehensive consideration of local user needs,adaptation to the surrounding environment,and green economy principles,this thesis proposes an improved non-dominated ranking genetic algorithm with elite strategy to optimize the 5G base station distribution scheme and improve the 5G user experience.The main work of this thesis is as follows:(1)The 5G base station planning problem is studied,and an improved algorithm of non-dominated ranking genetic algorithm with elite strategy based on improved box-crossover operator and rank-based t distribution variation operator is proposed based on the distribution rules of 5G base stations,and the effectiveness of the proposed algorithm is verified with five other related algorithms on four standard test functions.(2)A network model was constructed to evaluate the planning scheme to further validate the optimization effect of the improved algorithm.In this model,the variables encoded with real numbers are used to present the base station parameters configuration,and the radius of base station coverage provided by the urban macro-cellular cell radio propagation model is used as the range of values for the variables,and the user coverage level and station construction cost are selected as the two objectives of the multi-objective optimization algorithm.Finally,experiments show that the proposed algorithm in this thesis has better convergence and distributivity compared with other related algorithms.The application in5 G network planning also proves the superiority of the proposed algorithm in solving 5G base station distribution planning problem and can provide a better distribution scheme.