Optimizing NOMA Schemes for Future Communication Networks: Heuristic Algorithms-based Approaches

Technological advancements and emerging applications in the area of wireless networks, in particular in envisioned beyond 5G (B5G) networks, have introduced new design issues and corresponding optimization problems. However, as the network parameters gets larger, as in future heterogeneous dense wireless networks with the Internet of things (IoT), and large-scale sensory applications, these problems turn out to be computationally NP-Hard and require iterative nondeterministic algorithms to solve. Tremendous amount of computational cost can be saved by modeling such problems as discrete optimization problems, identifying the constraints, and solving them iteratively using modern heuristics. Among such optimization problems is the optimization of non-orthogonal multiple access (NOMA) and/or hybrid NOMA-orthogonal multiple access (OMA) schemes which were introduced to address the efficient sharing of resources in such dense networks, as well as the fairness aspects among the served users and their low latency requirements [13]. The optimization of NOMA-enabled schemes typically involves the clustering of the served users to allow ranking and SIC detection, the power allocation among them, and, possibly, the required channel assignment for hybrid NOMA-OMA schemes. The main aim of this proposal is to focus on investigating some NP-hard optimization problems in NOMA-enabled schemes for future communication networks and their solutions using efficient heuristic algorithms. Examples of such problems include channel allocation, power allocation, and user clustering [46]. Heuristics to be experimented with and engineered to solve the identified problems include both classical techniques such and Simulated Annealing, Tabu Search, Genetic Algorithms, [7], etc., and the more recent techniques which may include swarm based heuristics such as grasshopper search, particle swarm optimization (PSO), artificial been colony (ABC), etc., [8, 9] to name a few. The suggested methodology comprises: (a) An extensive literature review will be conducted and NP-hard combinatorial optimization problems will be identified. For example, the novel system model for NOMA-enabled mobile edge computing applications will be formulated and the associated optimization problem for objectives such as maximum throughput or maximization of spectral and/or energy efficiencies will be formulated. (b) Then, modern heuristics or metaheuristics will be engineered to solve the identified hard optimization problems. Further, in the second phase of the project, the candidate NOMA-enabled schemes for high data rate visible light communications (VLC) future applications will be identified and optimized subject to constraints on the signal design in these networks, with the objective to maximize the achievable rates and/or the energy efficiency. In both phases, the performance-complexity trade-offs will be investigated and the heuristics-based solutions will be compared to the optimal exhaustive search methods or with published benchmarks, where ever possible.