Model-Free Learning of Optimal Beamformers for Passive IRS-Assisted Sumrate Maximization

TL;DR

This paper introduces a novel data-driven Zeroth-order Stochastic Gradient Ascent algorithm for optimizing IRS beamformers in wireless networks, eliminating the need for channel models and outperforming traditional methods.

Contribution

The paper proposes a model-free, sensing-free algorithm for IRS beamforming that jointly optimizes long-term and short-term parameters without channel information.

Findings

ZoSGA outperforms model-based baselines in simulations.

Supports convergence analysis with state-of-the-art results.

Effective in various IRS-assisted wireless scenarios.

Abstract

Although Intelligent Reflective Surfaces (IRSs) are a cost-effective technology promising high spectral efficiency in future wireless networks, obtaining optimal IRS beamformers is a challenging problem with several practical limitations. Assuming fully-passive, sensing-free IRS operation, we introduce a new data-driven Zeroth-order Stochastic Gradient Ascent (ZoSGA) algorithm for sumrate optimization in an IRS-aided downlink setting. ZoSGA does not require access to channel model or network structure information, and enables learning of optimal long-term IRS beamformers jointly with standard short-term precoding, based only on conventional effective channel state information. Supported by state-of-the-art (SOTA) convergence analysis, detailed simulations confirm that ZoSGA exhibits SOTA empirical behavior as well, consistently outperforming standard fully model-based baselines, in a variety of scenarios.