Research On Network Selection Algorithm Based On Intelligent Method In Ultra Dense Heterogeneous Network

With the development of wireless communication technologies,now a number of new technologies have emerged.In particular,the commercial use of the fifth generation communication technology makes the next generation wireless network access technology more complex.5G and the existing communication technology together form a super dense heterogeneous wireless network that supports the terminal to enjoy seamless network services.In ultra dense heterogeneous wireless network environment,The end users choose the network according to their own and the surrounding environment,which has become the focus of current academic research.The current network selection algorithm based on artificial intelligence algorithm improves the adaptive ability of the system.However,in the ultra dense heterogeneous wireless network environment,its dynamic greatly increases,then it affects the system performance and user experience.In order to improve the performance of network selection system,this paper proposes a network selection algorithm based on improved deep Q-learning.Secondly,considering the mobility of terminal,this paper proposes a network selection algorithm based on improved neural network.Both of them consider throughput and network switching times to improve the overall performance of the system,and optimize the complexity of the algorithm and reduce the time cost.The main work of this paper is as follows:1.In ultra dense heterogeneous wireless network environment,aiming at 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 Q-learning 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,by using the transferred training samples and weights to accelerate the training of neural network,then we can get the optimal network selection strategy.Experimental results demonstrate that the algorithm significantly improves the performance degradation of high dynamic network handoff caused by sleep mechanism and the time complexity of traditional deep Qlearning algorithm for online network selection.2.In order to solve the problem of unnecessary handoff and network selection performance degradation that caused by the mobility of the terminal further aggravates the high dynamic of the network environment,we propose a network selection algorithm based on improved neural network.Firstly,the throughput of the terminal is calculated by predicting the service duration of the terminal in the connected network,which is defined as the network revenue.Secondly,the network decision is triggered according to whether the network revenue is less than the threshold value,so as to alleviate the frequent handoff problem caused by the high dynamic of the network.Finally,using the neural network optimized by whale algorithm,the received signal strength and dormancy probability are taken as inputs,and the network is selected according to the trained output value.Experimental results show that the proposed algorithm can reduce the number of terminal handoff in high dynamic network environment,improve the terminal satisfaction,and the time cost of the algorithm is reduced.