Task offloading is a key strategy in Fog Computing (FC). The definition of resource-constrained devices no longer applies to sensors and Internet of Things (IoT) embedded system devices alone. Smart and mobile units can also be viewed as resource-constrained devices if the power, cloud applications, and data cloud are included in the set of required resources. In a cloud-fog-based architecture, a task instance running on an end device may need to be offloaded to a fog node to complete its execution. However, in a busy network, a second offloading decision is required when the fog node becomes overloaded. The possibility of offloading a task, for the second time, to a fog or a cloud node depends to a great extent on task importance, latency constraints, and required resources. This paper presents a dynamic service that determines which tasks can endure a second offloading. The task type, latency constraints, and amount of required resources are used to select the offloading destination node. This study proposes three heuristic offloading algorithms. Each algorithm targets a specific task type. An overloaded fog node can only issue one offloading request to execute one of these algorithms according to the task offloading priority. Offloading requests are sent to a Software Defined Networking (SDN) controller. The fog node and controller determine the number of offloaded tasks. Simulation results show that the average time required to select offloading nodes was improved by 33% when compared to the dynamic fog-to-fog offloading algorithm. The distribution of workload converges to a uniform distribution when offloading latency-sensitive non-urgent tasks. The lowest offloading priority is assigned to latency-sensitive tasks with hard deadlines. At least 70% of these tasks are offloaded to fog nodes that are one to three hops away from the overloaded node.

The 3GPP standard defines the requirements for next-generation wireless networks, with particular attention to Ultra-Reliable Low-Latency Communications (URLLC), critical for applications such as Unmanned Aerial Vehicles (UAVs). In this context, Non-Orthogonal Multiple Access (NOMA) has emerged as a promising technique to improve spectrum efficiency and user fairness by allowing multiple users to share the same frequency resources. However, optimizing key parameters–such as beamforming, rate allocation, and UAV trajectory–presents significant challenges due to the nonconvex nature of the problem, especially under stringent URLLC constraints. This paper proposes an advanced deep learning-driven approach to address the resulting complex optimization challenges. We formulate a downlink multiuser UAV, Rate-Splitting Multiple Access (RSMA), and Multiple Input Multiple Output (MIMO) system aimed at maximizing the achievable rate under stringent constraints, including URLLC quality-of-service (QoS), power budgets, rate allocations, and UAV trajectory limitations. Due to the highly nonconvex nature of the optimization problem, we introduce a novel distributed deep reinforcement learning (DRL) framework based on dual-agent deep deterministic policy gradient (DA-DDPG). The proposed framework leverages inception-inspired and deep unfolding architectures to improve feature extraction and convergence in beamforming and rate allocation. For UAV trajectory optimization, we design a dedicated actor-critic agent using a fully connected deep neural network (DNN), further enhanced through incremental learning. Simulation results validate the effectiveness of our approach, demonstrating significant performance gains over existing methods and confirming its potential for real-time URLLC in next-generation UAV communication networks.