Hypergraph-Based Evolutionary Game for UAV Sensing-Assisted User Association in Fully Decoupled Networks

Abstract

In fully decoupled networks (FDNs), optimizing joint uplink (UL) and downlink (DL) user equipment (UE) association under high mobility is challenging due to complex co-channel interferences and massive strategy spaces. Although uncrewed aerial vehicle (UAV) sensing enables necessary realtime environmental perception, deriving optimal association policies via conventional decentralized approaches (e.g., evolutionary games) suffers from prolonged convergence and the curse of dimensionality. To address these issues, we propose a joint framework integrating UAV sensing with a novel hypergraph-assisted evolutionary game mechanism to maximize the network sumrate. Specifically, we develop an adaptive hypergraph spectral clustering algorithm to accurately model the overlapping interference. This approach partitions the large-scale network into independent sub-populations, effectively reducing the localized strategy spaces. Driven by discrete replicator dynamics, UEs iteratively update their decisions to converge to an evolutionarily stable strategy (ESS). Simulation results verify that the proposed framework significantly outperforms traditional unpartitioned baselines in terms of total sum-rate and convergence efficiency.