Throughput-Fairness Trade-off Optimization for IRS NOMA Systems

Improving the overall throughput (i.e., sum rate) of a communication system while maintaining fairness between users has recently been identified as a crucial requirement for beyond fifth-generation communication systems. However, these performance metrics, i.e., sum rate and fairness, are conflicting. Specifically, maximizing the overall sum rate is achieved at the cost of per-user throughput degradation and vice versa. Such a conflicting nature has an undesirable impact on fairness between users, especially those with weaker channel conditions. To deal with such an issue, this paper proposes a multi-objective optimization (MOO) resource allocation technique for intelligent reflecting surfaces (IRS)-assisted multi-carrier non-orthogonal multiple access (NOMA) systems. The proposed MOO framework aims to balance overall throughput and fairness between users, where the fairness index (FI) is selected as a quantitative measure of fairness. Unlike single-objective optimization (SOO) problems, traditional approaches cannot solve the formulated MOO framework. Consequently, the weighted max-min (WMM) method is deployed to transform the MOO problem into a conventional SOO one, showing that the WMM method can achieve a set of Pareto-optimal solutions (i.e., dominant solutions). Consequently, we use the sequential convex approximation to evaluate the optimization parameters, namely the allocated power levels and the phase-reflecting coefficients of the IRS units. A set of simulation results is carried out to demonstrate the superiority of the proposed fairness-throughput resource allocation technique. In addition, the performance of the proposed WMM method is compared with a benchmark method, namely the weighted sum method (WSM).