1: \begin{abstract}

2: 	Reducing computational complexity is crucial in optimizing the phase shifts of Intelligent Reflecting Surface (IRS) systems since IRS-assisted communication systems are generally deployed with a large number of reflecting elements (REs). This letter proposes a low-complexity algorithm, designated as Dimension-wise Sinusoidal Maximization (DSM), to obtain the optimal IRS phase shifts that maximizes the sum capacity of a MIMO network. The algorithm exploits the fact that the objective function for the optimization problem is sinusoidal w.r.t. the phase shift of each RE. The numerical results show that DSM achieves {\color{black}near maximal sum-rate} and faster convergence speed than two other benchmark methods.

3: \end{abstract}

4: