Joint power allocation and deployment optimization for HAP-assisted NOMA-MEC system

High altitude platforms (HAPs) have been considered a promising paradigm to assist mobile edge computing (MEC) in wide-area internet of things (IoT) scenarios due to their wide coverage, large payload capacity, and excellent channel quality. In this paper, a HAP-assisted MEC system is investigated, where the HAP is deployed as an edge server to provide computation offloading services for IoT terminal devices (TDs) with limited local computing capability. Offloading is enabled by uplink and downlink communications between TDs and HAP in a clustered non-orthogonal multiple access (C-NOMA) manner. Considering potential economic benefits and expected environmental impact, we aim to minimize system energy consumption, subject to the constraints on maximum tolerable delay, transmit power budget, and uplink and downlink decoding power. To this end, an iterative algorithm based on relaxation and successive convex approximation methods is proposed, which efficiently solves this challenging non-convex problem by alternately optimizing power allocation and HAP deployment. Numerical results show that the proposed scheme significantly reduces the system energy consumption compared to the other benchmark scheme. It is also shown that the system energy consumption is indirectly effected by the decoding order in the uplink and downlink C-NOMA.