Research On Access Selection Of Heterogeneous Wireless Networks Based On Fuzzy Neural Network Theory

As the smart phone communication has been popularized to the old,middle and young generations,the huge network demand drives the rapid development of wireless network.However,the business needs of these users are complex,and the performance indicators of wireless networks that can meet different needs are different to some extent,so it is impossible to satisfy all users by relying on wireless network,as a result,the trend of integration and development of heterogeneous wireless networks has gradually formed.Access selection technology is the key for users to obtain high quality service anytime and anywhere in the environment of heterogeneous wireless networks,which makes access selection technology of heterogeneous wireless networks become an eye-catching research field on heterogeneous wireless networks.This thesis comprehensively sorts out the existing access selection algorithms in heterogeneous wireless networks,with elaborating the basic theories and ideas of the algorithms,analyzing the advantages and disadvantages of the algorithms,proposing an access selection algorithm based on DA-FNN and an access selection algorithm based on BDT-GWO-FNN in heterogeneous wireless networks.Although the existing access selection algorithms in heterogeneous wireless networks can take into account both users and networks,there are still some problems,such as inability to quantitatively process fuzzy information and lack of adaptive learning ability,an access selection algorithm in heterogeneous wireless networks based on DA-FNN is proposed in this thesis.By combining fuzzy logic theory with neural network,the algorithm can adjust parameters adaptively while processing fuzzy network attribute information.The dragonfly algorithm is used to optimize the membership function parameters of fuzzy neural network,so that the evaluation results of network are more accurate and the system performance is improved.In addition,when choosing a network,considering the influence of service type on network access selection results,the business types are divided into real-time business and non-real-time business in this paper,the preference values of users of the two kinds of services for each alternative network are set,and then the preference values are used to weight the alternative network scores output by the fuzzy neural network.Finally,the weighted values are used as the selection basis.The simulation results show that the algorithm can select the most suitable network for users,reduce the blocking rate while ensuring the throughput.Aiming at the problem of access selection in heterogeneous wireless networks when end users move fast,this paper proposes an access selection algorithm in heterogeneous wireless networks based on BDT-GWO-FNN.At each decision moment,multi-conditional Bayesian algorithm is used to filter the alternative networks to get the networks that meet the handoff conditions,and then the updating rules of the target network set are established to determine whether they are updated or not.Handoff decision is initiated when updating,otherwise,handoff decision is not initiated.After handoff decision is started,the network selection algorithm of gray Wolf optimization fuzzy neural network is used to select the alternative network in the updated handoff target network set.Since both gray Wolf algorithm and dragonfly algorithm are swarm intelligence optimization algorithms,gray Wolf optimization fuzzy neural network and dragonfly optimization fuzzy neural network have similar ideas.Both of them first build the fuzzy neural network model,and then use the optimization algorithm to optimize the membership function parameters of fuzzy neural network,so as to improve the system performance.Simulation results show that the proposed algorithm can improve throughput,reduce blocking rate and decrease the number of handovers.