STATISTICAL PERFORMANCE ANALYSIS OF NOMA-COOPERATIVE WIRELESS-POWERED MOBILE EDGE COMPUTING NETWORKS

Abstract

Wireless-powered mobile edge computing (WP-MEC) offers battery-free operation for IoT edge devices, but published evaluations typically report results for a single fixed channel, leaving it unclear whether reported gains are general or channel-specific. This paper presents a statistical performance analysis of a cooperative, non-orthogonal-multiple-access-based wireless-powered edge computing network, in which three devices harvest energy from an access point and cooperate to offload computation. Three objectives are pursued: statistically characterising system performance across many independent random channel conditions which is the first such analysis in this area; benchmarking the proposed scheme against four alternatives, including orthogonal-access and battery-powered designs; and establishing practical operating boundaries in power, coverage, and data-rate. Across 500 independent random channel realisations, access-point energy consumption is minimised using a globally optimal algorithm, with mean, standard deviation, and confidence intervals computed across all successful trials and compared against four alternatives under matched conditions. The system achieves a workable solution in 76.2% of trials, reduces access-point energy by an average of 73.0%, and delivers 83.8% more throughput than the orthogonal-access alternative on average, with an advantage never falling below 29.3% in any trial. A battery-powered design fails to meet task deadlines in every trial, while the proposed system never does, yielding an energy-efficiency advantage exceeding two orders of magnitude. These findings confirm that reported gains are a statistically consistent property of the design, not a channel-specific artefact, and translate into three deployment guidelines: a data-rate ceiling of per device, a coverage radius of approximately, and a quantitative transmit-power recommendation for a given fairness target.