The rapid growth of low-Earth-orbit (LEO)satellites has enabled integrated space-air-ground networks to provide seamless connectivity to mobile users. However, these networks face challenges such as physical layer security risks from line-of-sight channels and the energy constraints of high-altitude platforms (HAPs), necessitating solutions for secure communication and energy efficiency. In this work, we address the challenges of energy efficiency and secure communication in space-air-ground networks, which are becoming critical with the increasing deployment of LEO satellites to support high-mobility users. We propose a novel downlink architecture where high-altitude platforms (HAPs) assist the LEO satellite in serving ground users. To tackle the demands of secrecy energy efficiency (SEE) in this dynamic and complex network, we formulate a non-convex optimization problem that jointly considers HAP trajectory, user-HAP association, and beamforming. The problem’s non-convexity makes it computationally challenging to solve in polynomial time. To overcome these challenges, we introduce a generative artificial intelligence (GAI)-based deep reinforcement learning (DRL) framework, named Gen-DRL, which leverages generative adversarial networks to empower its agents. This framework dynamically predicts and adapts to changes in the space-air-ground network environment by optimizing key parameters such as channel states, HAP trajectories, user associations, and beamforming. Compared to conventional methods, the proposed Gen-DRL achieves significant improvements in SEE by effectively managing complex interdependencies among multiple agents and intelligently adapting to the network’s goals and constraints. Extensive simulation results demonstrate that Gen-DRL consistently outperforms existing state-of-the-art frameworks in terms of secrecy energy efficiency, robustness to dynamic user locations, and adaptability to varying network parameters. This work provides new insights into the design of secure and energy-efficient space-air-ground networks, highlighting the potential of GAI-based DRL for future communication systems.

In this work, we investigate content caching problem in unmanned aerial vehicle (UAV)-enabled networks where UAVs are allowed to store user requested content so as to enhance the content transmission performance of users. To predict user content request, we propose a bidirectional long short-term memory-based algorithm. Then, the content transmission delay of users is examined and the UAV deployment, content caching and resource allocation problem is formulated as an overall content fetching delay minimization problem. As the formulated problem is difficult to solve, we transform it into two subproblems which are solved iteratively. To tackle the UAV deployment and content caching subproblem, we first design a modified K-means-based clustering scheme and then propose a UAV deployment strategy by using quadratic transformation. To solve subcarrier and power allocation subproblem, we apply Lagrangian dual method to determine power allocation strategy and propose a Kuhn-Munkres algorithm-based subcarrier allocation strategy. Simulation results demonstrate the effectiveness of the proposed algorithms.

Caching the contents of unmanned aerial vehicles (UAVs) could significantly improve the content fetching performance of request users (RUs). In this paper, we study UAV trajectory design, content fetching, power allocation, and content placement problems in multi-UAV-aided networks, where multiple UAVs can transmit contents to the assigned RUs. To minimize the energy consumption of the system, we develop a constrained optimization problem that simultaneously designs UAV trajectory, power allocation, content fetching, and content placement. Since the original minimization problem is a mixed-integer nonlinear programming (MINLP) problem that is difficult to solve, the optimization problem was first transformed into a semi-Markov decision process (SMDP). Next, we developed a new technique to solve the joint optimization problem: option-based hierarchical deep reinforcement learning (OHDRL). We define UAV trajectory planning and power allocation as the low-level action space and content placement and content fetching as the high-level option space. Stochastic optimization can be handled using this strategy, where the agent makes high-level option selections, and the action is carried out at a low level based on the chosen option’s policy. When comparing the proposed approach to the current technique, the numerical results show that it can produce more consistent learning performance and reduced energy consumption.

In this work, we investigate statistical quality-of-service (QoS) provisioning in a downlink reconfigurable intelligent surface (RIS)-aided holographic multiple-input multiple-output (MIMO) network. Holographic MIMO, enabled by nearly continuous aperture surfaces, offers fine-grained beamforming capabilities that can be leveraged for delay-sensitive and reliable communication. We aim to optimize discrete configurations of base station precoding, RIS phase shifts, and user-side combiners to meet statistical delay and reliability constraints. To this end, we propose a novel discrete softmax policy gradient (DSPG) framework that learns discrete control policies under effective capacity-based statistical QoS constraints. To enhance policy expressiveness and decision quality in large discrete action spaces, we integrate a generative diffusion model (GDM) into the agent architecture, enabling efficient sampling of high-quality configurations. Simulation results demonstrate that the proposed GDM-based DSPG significantly outperforms existing baselines, particularly under stringent latency and reliability requirements, making it a strong candidate for next-generation intelligent wireless systems.

The integration of uncrewed aerial vehicles (UAVs) and low Earth orbit (LEO) satellites is expected to provides flexible, scalable and efficient computing support for Internet of things (IoT) applications. In this paper, we study the task execution problem in a UAV-assisted satellite offloading network, where UAVs fly over the network coverage region to gather tasks from IoT devices. The collected tasks can be executed at the UAVs or offloaded to satellites for processing. Jointly considering the energy consumption and task dropping cost, we define system cost function as the weighted sum of the two metrics, and formulate a constrained system cost minimization problem, where the power allocation strategy, task offloading and scheduling, and UAV flight trajectory planning is jointly optimized. As the formulated problem is a mixed integer nonconvex optimization problem, which cannot be solved conveniently, we decompose it into subproblems, and propose an alternate iteration-based algorithm to solve the subproblems. In particular, to tackle the IoT device task transmission subproblem, we relax binary variables into continuous ones and transform the original problem into a standard linear problem which can be addressed using existing optimization tools. To resolve UAV trajectory design subproblem, we employ successive convex optimization and first order Taylor expansion method to convert the problem into a convex problem which can then be solved conveniently. For the power allocation subproblem, an iterative method and Lagrange dual approach are applied to obtain the optimal result. To obtain task offloading and computing scheduling strategy, a heuristic approach is proposed. The simulation results reveal that our proposed method achieves superior performance compared to the existing algorithms.

Energy efficiency (EE) is a critical requirement for next-generation wireless networks, motivating the use of rate-splitting multiple access (RSMA) and intelligent reflecting surfaces (IRSs) as low-power, interference-resilient technologies. This paper studies EE maximization in a UAV-mounted IRS-assisted multi-user MISO downlink under practical hardware impairments. A communication-centric EE metric is adopted, and the optimization of RSMA precoders, common-rate allocation, IRS phase shifts, and UAV placement is formulated as a non-convex problem. To solve it efficiently, we develop an alternating optimization framework based on successive convex approximation (SCA) and rank-one relaxation. Simulation results reveal that the proposed aerial IRS-assisted RSMA design achieves substantial EE gains over NOMA and SDMA baselines and remains robust to distortion, IRS size variations, and dynamic user conditions, highlighting its suitability for energy-constrained 6G deployments.