Irregular RIS-Assisted Cell-Free Massive MIMO Under Low-Resolution ADCs/DACs: ADHMA Approach

Abstract

This work investigates a multi-irregular reconfigurable intelligent surfaces (IRIS)-assisted cell-free massive multiple-input multiple-output system under practical hardware constraints. Specifically, we consider low-resolution analog-to-digital converters /digital-to-analog converters, discrete phaseshifts, spatial correlation among IRIS elements, and imperfect channel state information. Within this framework, we evaluate the downlink sum spectral efficiency (sum-SE) and formulate a joint optimization problem for the selection topology of active elements and the phase-shift matrix. To efficiently solve the resulting optimization problem, we propose an adaptive dual-Hamming memetic algorithm (ADHMA), which employs dual adaptive Hamming thresholds to maintain population diversity during global search and adaptively expand elite-based local search, achieving an effective balance between exploration and exploitation. Simulation results reveal that the proposed IRIS scheme achieves notable SE improvements over conventional RIS with same number of active elements. Furthermore, under practical hardware constraints, IRIS demonstrates more pronounced performance gains, achieving performance comparable to ideal hardware systems without RIS, while reducing the number of required active elements compared to conventional RIS for the same performance level.