Weighted sum rate optimization for intelligent reflecting surface-aided wireless network

For the transmission design problem in an IRS (intelligent reflecting surface)-enabled network, by jointly designing the transmit beamforming and IRS reflecting coefficient, the goal of this paper was to maximize the weighted sum rate for multiply ground users, subject to the transmit power and the unit modulus constraint. To solve the non-convex objective, we developed an alternating optimization method, where the phase shifter optimization was solved by the RMG (Riemannian manifold gradient) method, and the beamforming was obtained by the bisection search method. Furthermore, an element-wise block coordinate descent-based method was proposed to reduce the complexity of the RMG method. Simulation results verify the effectiveness of the proposed algorithm, and demonstrate that IRS can significantly improve the spectrum efficiency, when the reflecting coefficients are properly optimized.