This letter investigates a hybrid non-orthogonal multiple access (H-NOMA) based Unmanned aerial vehicle (UAV) assisted mobile edge computing (MEC) network. The goal is to reduce the offloading energy consumption in data-intensive and time-critical smart devices (SDs). To achieve this, we proposed latency-based clustering and low-complexity resource allocation algorithms. Instead of traditional clustering methods, we have introduced a latency condition where H-NOMA surpasses both NOMA and orthogonal multiple access (OMA) in effectively clustering the users. Furthermore, we derived and analyzed a closed-form resource allocation solution and compared it with OMA and NOMA. In addition, the simulation results indicate that the system’s performance utilizing the proposed algorithms outperforms NOMA and OMA approaches, whether with heuristic clustering or without clustering, thereby emphasizing its undeniable superiority.

In this letter, we present a novel method of solving the Generalized Waterfilling problem (GWFP) using equivalence with a waterfilling problem (WFP) which has a closed form solution. Consequentially, a class of WFPs with only one-step called `1-step' WFP is used to derive a low complexity algorithm which has the smallest worst case complexity reported. The proposed algorithm is verified by simulations.

Waterfilling problems subjected to peak power constraints are solved, which are known as cave-filling problems (CFP). The proposed algorithm finds both the optimum number of positive powers and the number of resources that are assigned the peak power before finding the specific powers to be assigned. The proposed solution is non-iterative and results in a computational complexity, which is of the order of M, O(M), where M is the total number of resources, which is significantly lower than that of the existing algorithms given by an order of M ² , O(M ² ), under the same memory requirement and sorted parameters. The algorithm is then generalized both to weighted CFP (WCFP) and WCFP requiring the minimum power. These extensions also result in a computational complexity of the order of M, O(M). Finally, simulation results corroborating the analysis are presented.

In this letter, we explore the hybrid Non-orthogonal multiple access (H-NOMA) premised Unmanned aerial vehicle (UAV)-Enabled mobile edge computing (MEC) network. Our research emphasizes the importance of adept clustering techniques to meet the high demands of data-intensive and time-sensitive users. Instead of conventional single-stage clustering, we introduce a novel 2-tier clustering mechanism. We assort users based on channel gains and cluster sizes in the initial phase. Then, these users undergo a second clustering based on their latency demands. This layered approach enables more efficient resource allocation tailored to individual user needs. Furthermore, we derive and analyze a formula for resource assignment, ensuring a more streamlined and efficient process. We compare our findings with pure NOMA, and orthogonal multiplexing access (OMA) benchmarks for clearer understanding. Simulations show the superior performance of the H-NOMA using our proposed methodologies.