Abstract:

Intelligent reflecting surface (IRS) can achieve unprecedented channel capacity gains by reconfiguring the wireless propagation environment smartly. The channel capacity is maximized by jointly optimizing the power-constrained precoder at the base station and the unit modulus-constrained phase shifter at the IRS in the IRS-assisted multi-user downlink communication. Aiming at the resulting non-deterministic polynomial hard problem, it is converted into an equivalent problem firstly, then the alternative optimization algorithm is used to solve the precoding matrix and phase shift vector. The problem can be converted to a second-order cone programming problem that can be solved by adopting standard optimization packages while the phase shifts vector fixing. The unit modulus constraint is the difficulty to solve this problem, it is decided to embed it into the search space. Then the Riemannian trust-region (RTR) algorithm is proposed to solve this problem while fixing the precoding matrix. Simulation results show that, compared with the existing approaches, RTR algorithm not only has better performance gains but also faster convergence speed.

Key words: intelligent reflecting surface (IRS), multi-user, channel capacity, unit modulus constraint, Riemannian manifold optimization