Physics-based Modeling of Large Intelligent Reflecting Surfaces for\n Scalable Optimization

In this paper, we develop a physics-based model that allows a scalable\noptimization of large intelligent reflecting surfaces (IRSs). The basic idea is\nto partition the IRS unit cells into several subsets, referred to as tiles, and\nmodel the impact of each tile on the wireless channel. Borrowing concepts from\nthe radar literature, we model each tile as an anomalous reflector, and derive\nits impact on the wireless channel for given unit cell phase shifts by solving\nthe corresponding integral equations for the electric and magnetic vector\nfields. Based on this model, one can design the phase shifts of the unit cells\nof a tile offline for the support of several transmission modes and then select\nthe best mode online for a given channel realization. Therefore, the number of\ntiles and transmission modes in the proposed model are design parameters that\ncan be adjusted to trade performance for complexity.\n