| In the era of rapid development of mobile communication technology,heterogeneous network is composed of wireless LAN,cellular network and metropolitan area network.In order to solve the contradiction between mobile users’ increasingly high requirements for communication quality and the scarcity of communication resources,it is necessary to require a variety of access networks to be integrated to ensure users’ communication service experience.However,because different types of communication systems have great differences in coverage,transmission rate and other aspects,it leads to grave difficulties in heterogeneous network convergence,among which handover management technology is an effective means to achieve heterogeneous network convergence.Therefore,making reasonable handover strategy plays an important role in heterogeneous convergence of multi-networks and is also the inevitable trend of the development of the next generation communication technology.Starting from the networking mode of heterogeneous network,this paper studies the problem of handover management in different heterogeneous network environments:1)In the current 4G network environment,in the ultra-dense heterogeneous cellular network composed of multiple 5G small cells densely deployed,due to the small base station coverage is small,mobile users will frequently switch.However,ping-pong effect and abnormal handover are caused by the unreasonable setting of handover parameters.Aiming at this problem,a handover strategy is proposed based on improved particle swarm optimization(PSO)algorithm to optimize handover parameters.The probability of radio link failures(RLFP)and ping-ping handovers(HOPP)can be effectively reduced by optimizing the handover parameters under different mobile speeds of users.Through a large number of experiments,the proposed algorithm can adaptively adjust the handover parameters according to different mobile speeds of users,so as to trigger the handover at the appropriate time,which not only reduces the ping-pong effect,but also effectively reduces the RLFP caused by premature,late and wrong handover.2)For the dense heterogeneous network composed of 5G heterogeneous cellular network and a variety of different types of wireless access technologies,the communication performance of different access technologies varies greatly.We study when and how mobile users switch to the appropriate network and propose a handover strategy based on mobile user location prediction and multi-attribute decision making.In the handover trigger stage,the hyperparameters set by long shortterm memory networks(LSTM)based Bayesian optimization are adopted(BL).In order to make the neural network to train and learn the user’s movement behavior rule faster and better,so as to build the user’s movement model.Then,the user’s location prediction model is loaded online to predict the user’s moving position at the next moment and estimate the network state at the future moment.When the handover trigger condition is met,the handover trigger mechanism can be activated in advance to reserve more processing time for the subsequent handover decision process,and realize the effect of reducing handover delay and unnecessary handover.In the handover decision stage,the unqualified candidate networks are screened out by the candidate network pre-selection mechanism,so as to reduce unnecessary handover and reduce the calculation time cost for the subsequent handover decision process.If the number of candidate networks is not unique,we use the multi-attribute decision mechanism to calculate the utility value of candidate networks by weighing users’ preferences for network parameters of different business types and the actual status of candidate networks,and take the candidate network with the largest utility value as the target network to complete network handover.According to the simulation experiment,the handover strategy proposed in this paper can significantly reduce the number of unnecessary handover and wrong handover times,and also reduce the handover delay.In addition,the benefits of network load balancing can be obtained while handover network reasonably for users. |