Efficient resource allocation in hardware-impaired cell-free systems is critical for achieving the stringent requirements of ultra-reliable low-latency communication (URLLC) while maintaining energy efficiency (EE). Traditional optimization-based approaches face scalability issues, while existing learning-based methods struggle to adapt to varying network structures and the complexities of hardware impairments. In this work, we address these challenges by incorporating rate-splitting multiple access (RSMA) into cell-free systems and designing a graph neural network (GNN)-based framework. First, we propose a method leveraging explicit channel state information to optimize precoding for both common and private streams under rate, power, and latency constraints. Next, we further develop an end-to-end approach that bypasses channel estimation by directly using raw pilot signals for joint feature extraction and optimization. Finally, we introduce a pilot-free method that processes distorted message-passing information from real channels, reducing communications overhead while enhancing adaptability to practical conditions. Through extensive simulations, we validate the proposed methods, demonstrating significant improvements in EE, along with insights into their computational complexity and scalability in diverse system configurations.

Future wireless networks aim to deliver ultra-reliable and low-latency services while containing their rapidly growing energy footprint. In a cell-free radio access network (CF-RAN), this objective translates into a tightly coupled optimisation over access point (AP) activation, user association, precoding design and virtual-CPU provisioning, all under finite-blocklength reliability constraints. We build a detailed power model that includes radio hardware, fronthaul and load-dependent computing, then recast the resulting energy efficiency problem as a mixed-integer second-order cone programming using a tight surrogate for decoding-error probability. A sparsity-promoting convex–concave solver can reach near-optimal solutions but must be run for every channel realisation, making real-time use impractical. To overcome this limitation, we propose a graph neural network (GNN) that represents CF-RAN as a heterogeneous AP-to-user graph, predicts precoding vectors, rates and soft association probabilities in a single forward pass, and then applies a lightweight reliability-enforcement layer to remove any residual violations. Simulation results show that, whenever the constraints are feasible, the learned solver achieves comparable energy efficiency as the optimization-based baseline while operating with only a single-pass inference step per channel realization.

This work explores the maximization of energy efficiency (EE) in cell-free ultra-reliable low-latency communication (URLLC) systems, specifically addressing the challenges presented by hardware impairments (HWIs). We introduce ratesplitting multiple access (RSMA) as an effective technique to enhance EE by managing the distortions caused by HWIs during downlink transmission. The optimization problem is formulated to maximize EE by optimizing precoding vectors for both common and private streams, while adhering to rate, power, and URLLC constraints. To address the non-convex and dynamic nature of this optimization problem, we propose a graph neural network (GNN) model that facilitates scalable and data-efficient solutions. Simulation results demonstrate that the proposed RSMA-GNN method consistently outperforms baseline approaches, including space-division multiple access (SDMA) and successive convex approximation (SCA) methods, particularly in scenarios characterized by severe HWIs.

In this paper, we investigate a cooperative edge caching problem in the fog radio access networks (F-RANs). In order to obtain the globally optimal caching strategy that minimizes the content transmission delay and maximizes communication efficiency, we propose a multi-agent reinforcement learning based cooperative caching policy with low entropy communications. First, we propose a double deep Q network (DDQN) based caching policy by taking into account the non-deterministic polynomial hard (NP-hard) aspect of this cooperative caching optimization problem. Then, we extend the state transition model of Markov Decision Process (MDP) under the single agent system into the Stochastic Game (SG) one under the multi-agent system. By employing the DDQN in each agent, the agents can learn and make the global decision for caching. For utilizing the cooperation resources of fog access points (F-APs), the interaction of information is introduced to exchange the historical cache records of cooperative F-APs. However, the information in the interaction may require lower entropy in the fiber link. Therefore, the information entropy is largely reduced to improve the communication efficiency by quantifying the information. Finally, due to the non-computable gradient of information entropy, we apply a pseudo gradient descent method to approximate the gradient descent in the local model. Simulation results show that our policy achieves better performance in terms of reducing the transmission delay and improving the cooperation among F-APs compared to the benchmark policies. Additionally, it is demonstrated that the proposed policy improves communication efficiency without compromising the performance of cooperative caching.

Ultra-Reliable and Low-Latency Communications (URLLC) is essential for sixth generation communications, with Cell-Free massive Multiple-input-Multiple-Output (CF mMIMO) being a promising architecture to support these demands. This paper addresses the challenge of optimizing energy efficiency in CF mMIMO systems for URLLC, focusing on the probabilistic delay bounds and finite blocklength communications. We propose a theoretical framework that considers tail distributions to evaluate extreme reliability and latency requirements, instead of relying on asymptotic analysis. In particular, a closed-form expression for the signal-to-interference-plus-noise ratio (SINR) distribution is derived, accommodating imperfections in channel state information caused by pilot contamination. Then, the paper also presents a comprehensive reliability analysis, incorporating both delay violation probability and average decoding error probability, utilizing stochastic network calculus for accurate statistical modeling. Finally, an innovative power control algorithm is proposed to maximize effective energy efficiency (EEE), the ratio of the effective data rate to total power consumption, while meeting stringent Quality-of-Service (QoS) constraints and power limits. Extensive simulations validate the theoretical framework and the efficacy of the proposed algorithm, demonstrating its ability to enhance EEE in various scenarios and providing insights into the interplay between EEE, delay, and reliability metrics.

In this paper, we investigate the energy efficiency (EE) optimization in the user-centric Cell-Free massive MIMO (CF mMIMO) system for Ultra-Reliable Low-Latency Communications (URLLC), where access points (APs) use maximum ratio transmission for downlink transmission. We first formulate an optimization problem to maximize the system EE while taking the finite blocklength achievable rate in URLLC into consideration. To deal with the intractable achievable rate in the objective function and the constraint, we derive a convex lower bound of it using successive convex approximation (SCA), and then reformulate the original problem into a second-order cone programming (SOCP). Next, we propose a low-complexity iterative algorithm to solve the SOCP by applying SCA. Simulation results show that the proposed method provides near-optimal performance in terms of Branch-and-Bound (BnB), and provide insights into the influences of blocklength, system parameters and AP clustering schemes on system EE.

As the demand for mobile data traffic continues to grow, offloading data traffic to unlicensed spectrum is a promising approach that can relieve the pressure on cellular systems. Therefore, it is an urgent need to propose an unlicensed spectrum access method to guarantee the harmonious and efficient coexistence between cellular network technologies such as LTE and incumbent users such as WiFi in the unlicensed spectrum. However, existing coexistence schemes such as licensed assisted access (LAA) and LTE-unlicensed (LTE-U) still suffer from inefficient spectrum utilization and unsatisfactory fairness. In the paper, we formulate the optimization problem of the unlicensed spectrum access among multiple small bases (SBSs) as a game, and then solve the Nash Equilibrium (NE) with cooperative and distributed multi-agent deep reinforcement learning (MADRL). Specifically, a two level access framework for the coexistence scenario, which consists of feedback cycle and executive cycle, is first proposed, and then the key elements of MADRL including state, action, reward and Q-network are designed in detail based on the proposed access framework. To overcome the problems of learning divergence and prohibitive computation overhead in the coexistence scenario with multiple SBSs due to the non-stability phenomena, we adopt the mean field technology to solve the NE, which can simplify the process of solving NE by converting the interaction of an agent with the remaining multiple agents into an action with the average effect of them. Simulation results show that 1) the proposed algorithm can overcome the learning divergence problem and converge to the NE quickly, and 2) the proposed algorithm can achieve the bi-objective optimization of total throughput and fairness of the coexistence network, and can achieve better performance in terms of throughput and fairness compared with the baseline methods such as Cat-4 LBT, Cooperative LBT and Random schemes.