Research And Design Of Heterogeneous Wireless Network Selection Algorithms

With the rapid development of mobile communication technology and broadband wireless access technology,various wireless communication technologies emerge in endlessly.Different technologies have differences in network performance indicators,supporting service features and other aspects,and a single communication technology cannot be suitable for various intricate application scenarios.Therefore,the next generation wireless network must be a heterogeneous wireless network integrated by a variety of different communication technologies.How to select the best network for user terminals in the heterogeneous environment has become a hot topic in the research of heterogeneous wireless network.Aiming at the shortcomings of some existing network selection algorithms,this thesis proposes three new heterogeneous wireless network selection algorithms from the perspective of algorithm improvement.The main contents are as follows:(1)An E-TOPSIS network selection algorithm based on the utility function is proposed to improve the inadequacies of E-TOPSIS algorithm in attribute normalization.Unlike E-TOPSIS,which uses the conventional standardization method,this improved algorithm realizes standardization through the sigmoid utility function,which not only flexibly reflects the needs of different attributes under various services,but also improves the ranking quality of candidate networks.The new algorithm first uses utility functions to evaluate utility and normalize,then uses M-AHP to calculate the subjective weight of network attributes,and finally selects the best network through E-TOPSIS.The simulation results show that,compared with the TOPSIS and E-TOPSIS algorithms,the new algorithm can significantly improve the abnormality of network rankings,and it can also effectively reduce the average number of handovers.(2)In order to improve the shortcomings of the R-GRA algorithm in reflecting the relational degree between the candidate networks and the reference networks,a network selection algorithm incorporating TOPSIS’s grey relational analysis method is proposed.The algorithm integrates the TOPSIS idea to comprehensively reflect the degree of association between each candidate network and the positive and negative reference networks in terms of distance and shape.In order to give consideration to the user’s own preference and the actual situation of the candidate networks,the new algorithm uses the subjective and objective comprehensive weighting method of the analytic hierarchy process and the entropy method to obtain the composite weight of all network attributes under each type of service,and uses the improved grey correlation analysis method to select the best network.The simulation results show that when compared with the GRA and R-GRA algorithms,the new algorithm can effectively reduce the average number of network handovers;when compared with the AHP-TOPSIS algorithm,the new algorithm can reasonably allocate network resources and maintain network load balance.(3)In order to improve the defect of network selection algorithm based on service priority when calculating the overall network parameter weight of service group,a multi-service network selection algorithm based on group decision is proposed.The new algorithm first obtains the composite weight of all network attributes under each type of service through the subjective and objective comprehensive weighting method of the analytic hierarchy process and the entropy method,and then multiplies the priority of each service by the composite weight vector of the corresponding service,and group decision-making is again used to synthesize these weight vectors,and again uses the capacitive criterion to judge the rationality of the weight vector synthesized by the group decision theory,which can ensure that the user cannot rely on their own preferences to choose some poorly performing networks,and finally use the TOPSIS algorithm to weight and sort to get the best network.The simulation results show that when compared with the network selection algorithm based on service priority,the new algorithm can better reduce the ping-pong effect of the network;when compared with the AHP-TOPSIS algorithm based on single service,the new algorithm can take into account the QoS requirements of multiple services,so as to efficiently and accurately select the best network.