Research On Network Selection Algorithm Of Heterogeneous Wireless Networks Base On Utility Function

With the rapid development of wireless communication technology,different types of access technologies have been existed,and the wireless network of next generation is a heterogeneous wireless network.Nowadays,the field of heterogeneous wireless networks has become the research hotspot and many research organizations and institutions have carried out the research on the related key technology of heterogeneous wireless networks.Network Selection is one of the key technologies of mobility management in heterogeneous wireless networks,to ensure the continuity of the session when users are roaming in the heterogeneous network.There are many network selection algorithms at present which have defects.From the improvement of the algorithm,two network selection algorithms based on utility function are proposed in this paper.By analyzing the characteristics of utility function in network selection,we construct the form of utility function in network selection.Through analyzing lack of additive utility and multiplying utility,mixed utility-based function network selection(MUFNS)algorithm is proposed.we classify the attribute according to the assessment of network QoS and evaluation of user characteristics.The weight of the attribute is constructed by the analytic hierarchy process and entropy weight method.The experimental results show that MUFNS algorithm can select the most appropriate network according to the user's business characteristics,and the optimal network selected is the same as the AHP-TOPSIS algorithm,and effectively reduce the ping-pong effect in network selection.In order to improve the shortcomings of DIA algorithm in attribute normalization,a utilitybased DIA network selection(UDIANS)algorithm is proposed.Utility functions are used to represent the requirement of application and utility values can be regarded as normalization properties,so that attribute normalization isn't changed when the worst network is removed.UDIANS uses utility function to evaluate utility and normalize,uses ANP to calculate the weight of network attributes,and finally makes network sorting through DIA,so as to select the best network.The results show that UDIANS eliminates the ranking abnormality of network and reduces the number of network handover.