In order to further improve the spectral efficiency of intelligent reflecting surface (IRS)-assisted millimeter wave non-orthogonal multiple access (NOMA) system, a joint design scheme of user clustering, successive interference cancellation (SIC) decoding order, transmit power allocation, analog beam selection and IRS phase shifting was investigated, under the transmit power constraint, SIC decoding rate constraint, beam selection constraint, SIC decoding order constraint, user rate constraint, and IRS phase shift constraint, aiming at maximizing system sum rate.Because of the coupling of discrete variables, combinatorial variables and continuous variables in objective function and constraints, the formulated optimization problem was difficult to solve by traditional optimization methods.To solve it, first the SIC decoding sequence variables were transformed into binary discrete variables through a query table, then the discrete variables and continuous variables were decoupled based on the hybrid whale optimization algorithm, and delt with the optimization constraints based on the penalty function method, and the non-convex optimization problem was transformed into a heuristic problem for iterative solution.Simulation results verify the effectiveness of the proposed scheme.
