A Novel MADM based Network Interface Selection Approach with Rank Reversal Avoidance in HWNs

Network interface selection (NIS) aims at choosing the best connection of the user equipment (UE) to the best available wireless network at any time and everywhere, satisfying the user’s and/or the service requirements in terms of quality of service (QoS). Although multi-attribute decision-making approaches(MADM) are commonly used to solve NIS problem, they suffer from couple of weaknesses: i) the rank abnormality that changes the alternatives’ ranking after discovering new networks or removing an existing one from the HWN environment, and ii) they always select the alternative having the highest ranking score without considering the user’s and/or service requirements. In this paper, we introduce a novel approach based on combining the Combined Compromise Solution (COCOSO) approach with utility function, called COCOSO-Based Utility function (COCOSO-UF), to overcome the COCOSO limits. We remove the rank reversal problem and we rank available networks based on the user or/and application needs. The obtained experimental results showed that our approach can avoid the rank reversal problem and satisfy the user or/and application QoS requirements.