Adaptive Access Selection Algorithm for Multi-Service in 5G Heterogeneous Internet of Things

Abstract

5G heterogeneous network system incorporates multiple radio access technologies (RATs), which enables the massive connection of Internet of things (IoT) devices and popularity of diverse IoT services. However, with the tremendous growth of IoT connections, personalization of service requirements and deepening of network heterogeneity, how to adaptively optimize the quality of experience (QoE) of IoT users in any motion state while balancing network load is still a major challenge in 5G system. Therefore, this paper proposes an innovative access selection mechanism named REMNS that allows users requesting IoT services to obtain optimal QoE in 5G heterogeneous networks. In particular, a network pre-assessment mechanism based on fuzzy logic is exploited to filter available networks by user devices. Furthermore, REMNS devises a comprehensive preference evaluation framework based on the subjective-oriented Analytic Hierarchy Process (AHP) and objective-oriented Entropy Weight Method (EWM) to measure the preference degrees of each IoT service for network attributes. Subsequently, we put forward a relative entropy based multi-service access selection algorithm to make servers rapidly sort optimal network from the filtered available networks, so as to enhance QoE for users under the constraint of limited network capacities. The evaluation results demonstrate that the REMNS can effectively maintain stable service connections and significantly improve user QoE.