The swift proliferation of Internet of Things (IoT) applications necessitates wireless infrastructures that are scalable, energy-efficient, and able to satisfy rigorous quality-of-service (QoS) criteria. Traditional cellular networks frequently have difficulties in overcoming these obstacles due to restricted spectrum reuse, variable coverage, and elevated energy usage. Recently, unmanned aerial vehicles (UAVs) and intelligent reflecting surfaces (IRSs) have surfaced as potent alternatives for improving coverage and spectrum efficiency. This study presents a reinforcement learning (RL) optimization methodology for a cell-free UAV–IRS-supported IoT network. The suggested architecture enhances throughput, broadens coverage, and diminishes energy consumption through the optimization of UAV positioning, IRS phase modifications, and resource allocation. The efficacy of the proposed strategy is proved through a comparative analysis of its performance against multiple baseline methodologies. Comprehensive simulation findings indicate the superiority of the RL-based methodology, yielding a total throughput enhancement of 15.38% compared to the most effective baseline, augmenting coverage by 5.56%, and decreasing energy usage by 4.55%. The collaboration between UAVs and IRSs is emphasized, with UAVs offering flexible relaying and IRSs improving signal quality in resource-constrained settings. These findings confirm that reinforcement learning is an effective approach for managing next-generation UAV–IRS–cell-free IoT systems, providing substantial performance enhancements while maintaining sustainability and energy efficiency.
Wu et al. (Tue,) studied this question.