Multiple Objective Channel Allocation Problem in 5G Networks

Abstract

5G networks, as an emerging technology, calls for novel solutions to several research problems, many of them having in focus a better usage of sparse spectrum capacity. In the attempt to leverage spectrum, additional unlicensed band, available for wireless technology, is combined with the licensed part of the spectrum, managed by traditional mobile providers, thus making a heterogeneous network environment of Primary Networks (PNs). In such an environment desirous users (i.e. Secondary Users (SU)) equipped with multiple radio access technology are able to select the appropriate network considering not just interference, but other parameters as well (e.g. QoS, data rates, prices etc.). We consider the architecture with central Cognitive Network Provider (CNP) and several heterogeneous PNs, in which SUs are contending for empty channels in one of the PNs that best suit their needs for bandwidth, data rate and price. Applying cognitive radio network principles leads to the well-known problem of Network Selection and Channel Allocation. CNP has to solve this NP-hard problem by considering equally the demands of the heterogeneous PNs as well as the best interest of SUs. In this paper we propose two different bi-objective models and solve them by generating their non-dominated points. First model allocates SUs to networks with respect to costs, target interferences, and data rate capacities; while the second model splits PNs in channels, and allocates SUs to channels with respect to their additional demands for low latency. We generate the efficient sets of the instances found in the literature; and compare our results with the existing results obtained by the nature inspired meta-heuristic approaches.