基于惩罚性变权的动态网络接入选择算法  

A Dynamic Access Network Selection Algorithm Based on Punishment Variable Weight

作　　者：陈智雄[1] 崔娟平 张华乐 苑津莎[1] 韩东升[1] CHEN Zhi-xiong;CUI Juan-ping;ZHANG Hua-le;YUAN Jin-sha;HAN Dong-sheng(Electronics and Communication Engineering, North China Electric Power University, Baoding 071003, China;State Grid Anhui Information & Telecommunication Company, Hefei 230061, China)

机构地区：[1]华北电力大学电子与通信工程系,保定071003 [2]国网安徽省电力公司信息通信分公司,合肥230061

出　　处：《科学技术与工程》2018年第35期43-49,共7页Science Technology and Engineering

基　　金：国家自然科学基金(61601182;61771195);河北省自然科学基金(F2017502059;F2018502047);中央高校基本科研业务费专项资金(2017MS109)资助

摘　　要：针对常权评估导致的网络选择不均衡问题,提出了一种联合结构熵权和惩罚性变权的网络选择算法。首先根据用户和业务特点确定指标体系,并采用结构熵权法计算指标常权重。然后根据实际网络参数,通过惩罚变权函数对常权重进行动态调整。最后计算各备选网络的综合性能,确定最佳接入网络。该算法同时考虑了用户偏好及实际网络参数,使判决结果兼顾用户与网络双侧利益。仿真结果表明,网络负载较高时,所提算法可有效降低网络阻塞率、均衡网络间负载,保证用户与网络双侧收益。Considering the problem of network selection imbalance caused by the constant weight evaluation,a network selection algorithm based on structural entropy weight and variable weight was proposed.Firstly,the index system was established according to the users preference and the traffic requirement.The structure entropy weight algorithm was used to calculate the constant weight of the decision index.Then,according to the actual parameters of the network,the weight of the judgment index was dynamically adjusted using punishment variable weight function.Finally,the overall performance score of each alternative network was calculated and the network with maximum score was selected as the best network.The algorithm takes into account the users preferences and the actual network parameters,so that the outcome of the decision can satisfy the requirement of both users and the networks.The simulation results show that the algorithm can effectively reduce the network blocking rate,balance the load between the networks and ensure the user and the network revenue when the network load is high.

关 键 词：异构无线网络 接入网选择 性能评估 结构熵权 惩罚变权 

分 类 号：TN929.5[电子电信—通信与信息系统]