Adaptive Collaboration in Multiplex UAV Networks with Dynamic Communication

Abstract

In recent years, UAV cluster architectures have evolved into multiplex networked structures, where UAVs are organized into tightly coupled subnets according to task demands or spatial distribution. These subnets operate under heterogeneous environmental conditions and interference patterns, resulting in highly dynamic inter-network communication. Such variability introduces two fundamental challenges. First, time-varying delays cause temporal misalignment in shared state information and degrade decision timeliness. Second, fluctuating bandwidth limits inter-subnet connectivity, impairing global situational awareness and collaborative efficiency. To address these challenges, we propose the Temporal Hierarchical Attention Actor–Critic (THAAC) algorithm. THAAC leverages a hierarchical attention mechanism to capture intra- and inter-subnet features and employs temporal prediction to compensate for delayed or missing communication, thereby enhancing perceptual synchronization and strategy consistency across the cluster. Simulation results show that THAAC achieves superior collaborative performance in multiplex networked UAV settings, delivering higher cumulative rewards and reduced communication cost compared with baseline approaches, thus enabling more adaptive and efficient cluster-level coordination.