Irregular Element Selection for Intelligent Reflecting Surface With Mutual Coupling Using Multiobjective Genetic Algorithm

This paper investigates the irregular intelligent reflecting surface (IRS) topology with a focus on mitigating the mutual coupling effect on the IRS and aiming to improve the received signal of user equipment and IoT devices. A multi-objective genetic algorithm (MOGA)-based method is proposed to tackle the topology design and signal channel of an irregular IRS. The element topology of an irregular IRS is encoded into a chromosome, and the mechanism of crossover and mutation is designed to find the topology with lower mutual coupling and higher channel quality. To improve the efficiency of the designed MOGA, the probability of selecting elements is adjusted from high to low according to the main lobe direction from the base station (BS). Simulation results show that the achievable rate of users and IoT devices is improved by using the proposed MOGA compared to other heuristic selection strategies and existing works. In addition, the sum rate obtained by using a specific probability distribution is better at the later stage of the proposed MOGA, which validates that the probability modification based on BS’s main lobe has an improvement on MOGA efficiency and system performance for users and IoT devices.