Secrecy rate optimization for intelligent reflecting surface assisted PD-NOMA system

A fundamental issue in the non-orthogonal multiple access (NOMA) technique is the ranking of users for appropriate successive interference cancellation (SIC). This depends on the available user channel gain. However, in networks that use intelligent reflecting surfaces (IRS) for communication, continuous phases at the IRS are not feasible and channel gain is changed by adjusting the phase shift matrix. In this paper, our goal is to increase the sum secrecy rate (SSR) of the physical layer security in downlink by assuming the channel state information of all the network components to be known and also considering the requirements of power-domain NOMA (PD-NOMA) and IRS. To this end, we formulate a joint optimization problem and divide it into four sub-problems; resource allocation, user decoding order or ranking, power allocation, and phase shift matrix design. Since the proposed optimization problems are non-convex, the solution was carried out using a recursive algorithm known as the alternative search method and the mesh adaptive direct search algorithm. As well as the network performance is illustrated for resource allocation scenarios using the random method, exhaustive search algorithm, and the proposed method. Numerical results indicated that the proposed scheme considerably improved the SSR compared to the case when the power and subcarrier allocation is performed randomly and the case when the phase shift matrix is considered constant.