Application of Dempster-Shafer theory for network selection in heterogeneous wireless networks

Network selection in heterogeneous wireless networks aims to choose an optimal network from several candidate wireless networks according to different constraints presented by, for example, applications, subscribers or operators. It has been studied with different methodologies such as decision function (DF) and multi-attribute decision making (MADM). In this paper, the Dempster-Shafer theory (DST) is applied to coping with more complicated scenarios during network selection. The basic idea is to combine the network evaluations (i.e. degrees of belief) judged from the viewpoints of different constraint factors to arrive at an agreement with joint mass for each network. Thus a rank of these networks is obtained and the one with the largest joint mass in the rank will be selected as the best network. The key concepts and the detailed application process of DST for multi-objective decision making of network selection are illustrated. And then two extended applications of DST are discussed for more complicated scenarios during network selection. Simulation results verify the consistency and the effectiveness of the mechanism in contrast to the latest MADM approach.