基于聚类分析和深度学习的多频多模网络负载均衡优化

负载均衡问题是LTE多频多模网络要解决的重大问题。多频多模网络结构复杂,负载均衡涉及的参数达数百个,仅依靠人工经验很难进行精细化配置。为解决多频多模网络的负载均衡问题,解决现网运维的难点与痛点,提出一种基于机器学习的多频多模网络负载均衡方案。首先选取关键指标对网络场景进行划分,然后利用机器学习技术挖掘出不同场景下的最佳参数配置建议。经验证,机器学习技术可以大大提高参数配置的质量和效率,做到精细化参数配置。

Load balancing based on clustering analysis and deep learning for multi-frequency and multi-mode network

Load balancing is a huge challenge for LTE multi-frequency and multi-mode network.Hundreds of parameters are involved in load balancing for the complex network structure.Therefore,it is difficult to perform precise and meticulous configuration only relying on human experience.In order to cope with the challenge,a load balancing scheme based on clustering analysis and deep learning was proposed.Firstly,the key indicators were selected to identify the network scenes,and then big data and deep learning technologies were used to mine the relationship between data.Finally,the optimum system parameters for different network scenes were found.It has been proved that machine learning technology can greatly improve the accuracy and the efficiency of parameter configuration.