Low-complexity particle swarm optimisation-based adaptive user clustering for downlink non-orthogonal multiple access deployed for 5G systems

Non-orthogonal multiple access (NOMA) has been envisioned as a fundamental method towards fifth generation (5G) cellular networks. Typical clustering schemes employ adaptive user clustering (AUC) to improve the performance of the NOMA system using brute-force search (BF-S). But, the search to perform AUC is computationally complex and practically infeasible. Therefore, AUC using particle swarm optimisation (PSO) algorithm is proposed to minimise the computational complexity. PSO is an intellectual algorithm, implements using a random number of particles moving in a search space. The particles are evaluated by the fitness value on each iteration until it reaches the optimal solution. Simulation results demonstrate that NOMA system employing PSO-based AUC is able to reduce the complexity with acceptable throughput performance compared with BF-S-based AUC. Furthermore, it is noteworthy that the proposed PSO-based AUC outperforms the conventional clustering with fixed number of users in NOMA system and orthogonal multiple access (OMA) system in terms of throughput performance.