基于DDQN的长江新链网络切片资源分配算法

Yangtze river new chain network slice resource allocation algorithm based on DDQN

针对长江新链网络为航运业务提供网络资源时能耗较高的问题进行了研究,采用了网络资源和算力资源统一建模为底层网络资源的方法,构建了网络切片环境下网络资源管理模型。根据长江新链网络切片资源的特点,构建了最小化能耗的目标函数。通过构建马尔可夫决策模型,提出了基于双深度Q-Network(Double Deep Q-Network,DDQN)的长江新链网络切片资源分配算法。通过实验与相关算法进行了分析,验证了文中算法选择能耗低的网络资源为航运业务提供资源,降低了长江新链网络资源约13.7%的能耗,提升了约10.2%的资源分配成功率。

A study was conducted on the high energy consumption of the Yangtze River New Chain Network in providing network resources for shipping business.The method of unified modeling of network resources and computing power resources as underlying network resources was adopted,and a network resource management model was constructed in a slicing environment.Based on the characteristics of the network slicing resources,construct an objective function to minimize energy consumption.A resource allocation algorithm based on DDQN(Double Deep Q-Network)was proposed by constructing a Markov decision model.Through experiments and analysis of relevant algorithms,it has been verified that the algorithm in this paper selects low energy network resources to provide resources for shipping business,reducing the energy consumption of the network resources by about 13.7%and improving the resource allocation success rate by about 10.2%.