Amid the explosive growth of latency-aware and computation-sensitive services, mobile edge computing (MEC) assisted by aerial networks, such as high-altitude platforms (HAPs) and low-altitude unmanned aerial vehicles (UAVs), has emerged as an effective solution for providing computational capabilities to regions with sparse terrestrial infrastructure. Nevertheless, aerial networks are highly sensitive to energy cost and inherently constrained in hosting dense computing resources, while the exposed wireless environment renders them particularly vulnerable to attacks from malicious nodes. Consequently, it is imperative to develop effective task scheduling and computing resources management mechanisms that satisfy users quality-of-service (QoS) requirements while minimizing system cost and ensuring network reliability. In this paper, we develop a multi-cell MEC network composed of multiple UAVs and multiple HAPs, and further propose a dual-layer blockchain-enabled, federated election (FE)-based (DBFE) group relative policy optimization (GRPO) algorithm to jointly reduce the task offloading latency and system energy expenditure. In particular, blockchain techniques enhance system resilience against non-Byzantine failures, whereas the FE mechanism suppresses the influence of Byzantine behaviors. Simulation results demonstrate that, compared with existing approaches, the proposed method reduces the overall system cost by 19% and 31% under scenarios without malicious nodes and with malicious nodes, respectively.
Haoyu Wan, Meng Li, Qi Li et al.· IEEE Transactions on Cogniti...· 0 citations
User devices (UDs) with non-independent and identically distributed (non-IID) data will worsen accuracy performance of the global model in federated learning (FL). Therefore, the implementation of diffusion strategies in machine learning (ML) models can enhance the effectiveness of federated learning with non-IID data. However, in a device-to-device (D2D) wireless federated learning (WFL) system, limited wireless resources and severe wireless channel interference become the important bottleneck to restrict the diffusion performance and model aggregation so as the global model of WFL with non-IID suffers from the weight divergence challenge. Thus, we propose a novel joint over-the-air computation (OAC) aggregation and diffusion framework by using a graph neural network (GNN) for WFL, termed an OAC-GNN-Dif framework. By integrating the OAC with message passing neural network (MPNN) of GNN, we further develop the OAC-MPNN-Dif algorithm based on the OAC-GNN-Dif framework. To further reduce communication costs, we designed an OAC message recurrent neural network (OAC-MPRNN-Dif) algorithm, where each UD propagates local models via D2D communications to refresh the graph embedding in the current frame based on the graph feature extraction and localization state of the previous frame to reduce communication costs. Additionally, we introduce dynamic time-varying MPNN for federated diffusion within evolving D2D network topologies. The experimental results indicate that our approach significantly performs well in communication overhead, with a 30%-60% decreasing in wireless resources overhead and 1.2-3.5 times decreasing in the number of model transfers compared to the FedDif methods. Moreover, our approach also improves the global model test accuracy, which is about 2.7% higher than the existing communication diffusion FL with non-IID characteristics.
Yunli Ji, Jiechun Zheng, Hongyang Du et al.· IEEE Transactions on Mobile...· 0 citations