Resource Allocation in 5G/IoT Networks

Abstract

Optimal utilization of resources like base station (BS) selection, user clustering (UC), user association, power, and spectrum allocation are ambitious targets of future 5th generation (5G) networks. 5G network will require immense capacity due to rise in the number of multimedia applications and mobile devices. In this regard, heterogeneous networks (HetNets), when compared to the existing traditional single-tier homogeneous networks, can increase the network throughput to satisfy the drastic increase in capacity demands. Homogeneous networks include only a macro base station (MBS), where HetNets include MBS and several small base stations (SBSs) which result in high throughput and energy efficiency (EE) of the network. Another important aspect of 5G is user access network, which is based on multiple access (MA) schemes. In the previous generations, orthogonal multiple access (OMA) scheme was used in which only one subcarrier (SC) was allocated to a user in the near vicinity and remote location. A user was associated with a remote MBS due to high transmit power compared to the nearby SBS with low transmit power in 4th generation (4G) HetNet. This scheme is not able to fulfill vast growing demands of users for multimedia applications, entertainment purposes, and internet usage in 5G. Hence, to meet the demands of future 5G, a novel nonorthogonal multiple access (NOMA) scheme, in which more than one user can be allocated to a single SC can be explored. This scheme will introduce interference among users on a single SC that can be mitigated using successive interference cancellation (SIC). Also, UC can be explored to further mitigate inter-cluster interference. UC can be defined as the group of users assigned an orthogonal SC. Moreover, hybrid nonorthogonal multiple access (HNOMA) scheme will ensure sub-optimal and economical utilization of resources, i.e., infrastructure, power, and frequency in future 5G HetNets. H-NOMA is a combination of OMA and NOMA. H-NOMA can meet the increasing demands of multimedia applications and mobile devices in 5G HetNet. In this thesis, we first explore the application of NOMA in future 5G and beyond 5th generation (B5G) networks. Then UC-based downlink H-NOMA scheme is used to formulate throughput maximization and EE maximization problems in terms of user admission in clusters, user association with BSs, and power allocation in HetNets. The formulated optimization problems fall into the class of mixed integer nonlinear programming (MINLP) problems, and an outer approximation algorithm (OAA) is proposed as a sub-optimal solution with less computation complexity. Performance analysis of the UC-based downlink H-NOMA scheme in the MBS-only network and HetNet is conducted by employing OAA that ensures efficient resource allocation for throughput maximization and EE maximization.