Resource Allocation for Throughput Maximization in Satellite-Terrestrial Integrated Networks

Abstract

To meet the futuristic communications needs, a satellite-terrestrial integrated network (STIN) has been proposed and is a strong contender amongst emerging architectures. Different approaches have been researched and mainly two models are being considered for STIN implementation. In our STIN model, we have considered a satellitebased base station, dovetailed with a terrestrial N-tier heterogeneous network (HetNet). Joint admission control, user association (UA), and power allocation while ensuring fairness while associating user equipment (UE) in STIN and fairness in the allocation of spectrum resources to associated UEs in STIN with an objective to maximize throughput has not been investigated in the past. Classically, a macro base station (MBS) has the maximum resources as compared to small base stations and in HetNets a UE associates with a single BS depending upon the received signal strength. Consequently, most UEs are expected to be associated with a dominant transmit power MBS which is not an optimal approach as various new challenges arise such as unfair traffic load and interference resulting in overall reduced throughput. In the proposed approach, we have made an endeavor to meet these challenges and formulated a throughput maximization problem considering joint admission control, fair UA, power, and fair spectrum resource allocation. The formulated problem is a mixed integer nonlinear programming (MINLP) problem that is non-deterministic polynomial-time hard (NP-hard) and to achieve an optimal solution it requires exhaustive search. But, the computational load of exhaustive search increases exponentially as the number of UEs increases. Therefore, to obtain a near-optimal solution having low computational load an outer approximation algorithm (OAA) is proposed. To evaluate the proposed algorithm, extensive simulation work has been performed. The effectiveness of the proposed approach is verified by the results in terms of fairness in UA, fairness in resource block (RB) allocation, and throughput in the downlink (DL).