基于改进深度Q学习的网络选择算法  

Network Selection Algorithm Based on Improved Deep Q-Learning

作　　者：马彬[1,2] 陈海波 张超[1,2] MA Bin;CHEN Haibo;ZHANG Chao(Chongqing Key Laboratory of Computer Network and Communication Technology,Chongqing University of Post and Telecommunications,Chongqing 400065,China;Institute of Computer Science and Technology,Chongqing University of Post and Telecommunications,Chongqing 400065,China)

机构地区：[1]重庆邮电大学重庆市计算机网络与通信技术重点实验室,重庆400065 [2]重庆邮电大学计算机科学与技术学院,重庆400065

出　　处：《电子与信息学报》2022年第1期346-353,共8页Journal of Electronics & Information Technology

基　　金：重庆市教委科学技术研究重大项目(KJZD-M201900602);重庆市教委科学技术研究重点项目(KJZD-M201800603);重庆市基础研究与前沿探索项目(CSTC2018jcyjAX0432);重庆市研究生科研创新项目(CYS20256)。

摘　　要：在引入休眠机制的超密集异构无线网络中,针对网络动态性增强,导致切换性能下降的问题,该文提出一种基于改进深度Q学习的网络选择算法。首先,根据网络的动态性分析,构建深度Q学习选网模型;其次,将深度Q学习选网模型中线下训练模块的训练样本与权值,通过迁移学习,将其迁移到线上决策模块中;最后,利用迁移的训练样本及权值加速训练神经网络,得到最佳选网策略。实验结果表明,该文算法显著改善了因休眠机制导致的高动态性网络切换性能下降问题,同时降低了传统深度Q学习算法在线上选网过程中的时间复杂度。In ultra dense heterogeneous wireless network with sleep mechanism,in view of the problem that the network dynamic is enhanced and the handoff performance is reduced,a network selection algorithm based on improved deep Q-learning is proposed.Firstly,according to the dynamic analysis of the network,a deep Qlearning network selection model is constructed;Secondly,the training samples and weights of the offline training module in deep Q-learning network selection model,which are transferred to the online network decision-making module through the transfer learning;Finally,the training samples and weights of transfer are used to accelerate the process of training neural network,and the optimal network selection strategy is obtained.Experimental results demonstrate that the proposed algorithm improves significantly the performance degradation of high dynamic network handoff caused by sleep mechanism and the time complexity of traditional deep Q-learning algorithm for online network selection.

关 键 词：超密集异构无线网络 改进深度Q学习 网络选择 

分 类 号：TN915[电子电信—通信与信息系统]