Network Selection Algorithm Of Heterogeneous Wireless Networks Based On MADM

Dramatic progress in wireless communication technology has ushered in a new phase in the development of wireless communication technology,in which different heterogeneous wireless networks coexist and complement each other.However,none of the heterogeneous wireless networks suits the requirements of high bandwidth,low cost,low delay and wide coverage.At the same time,mobile terminals exhibit the ability to support different network access technologies.As a result,integration of heterogeneous wireless networks is becoming a definite trend.How to select the most suitable network from a complex network environment is a key question that should be resolved in the integration of heterogeneous wireless networks.This research has conducted an in-depth study in network selection algorithms based on multi?attribute decision making with respect to terminal battery,weighting and ranking method.Terminal battery constitutes a factor that should not be neglected in network selection,but most of the present network selection algorithms do not take into account the influence of battery consumption,or fail to provide an effective adaptive model.This research,therefore,presents a network selection algorithm based on the adaptive weights of terminal battery levels.The algorithm can automatically adjust the weights according to the percentage level of the terminal battery;the lower the terminal battery level,the greater the variation of its weight.Our simulation has shown that the algorithm is inclined to select the network that entails less energy consumption so as to effectively reduce network handover times and save energy.In network selection,AHP is a common algorithm of calculating the weight of attributes,but this algorithm is rather subjective,and it is prone to make ambiguous judgment when it constructs contrast matrix.This research proposes a weighting method based on Bayesian estimation,which uses the weight obtained by the Delphi method as prior information,and the weight generated by AHP as sample information.By Bayesian estimation,the sample information is employed to modify the prior information so as to obtain the posterior weight of attributes.Our simulation has shown that this algorithm can relieve the problem of neglecting user preference with Delphi algorithm,and moreover,reduce the influence of ambiguous judgment by AHP on the result of the decision.TOPSIS is the most widely used network selection algorithm.However,by this method it is easy to produce ping-pong effect,along with some other problems in the calculation of closeness.This research,therefore,proposes the Improved-TOPSIS algorithm,which works to revise the method of calculating closeness and use the history information of candidate networks,in combination with the utility function,to alleviate ping-pong effect.The simulation has proved that the accuracy of the Improved-TOPSIS is nearly eight percentage points higher than that of the TOPSIS,reducing the handover times of terminal network and the probability of unnecessary handover,and saving resources of the terminal and the network.