PAST: Pilot and Adaptive Orchestration for Timely and Resilient Service Delivery in Edge-Assisted UAV Networks under Spatio-Temporal Dynamics
Houyi Qi, Minghui Liwang, Liqun Fu, Sai Zou, Xinlei Yi, Wei Ni, Huaiyu Dai
TL;DR
PAST tackles resource trading in edge-assisted UAV networks with spatio-temporal dynamics by fusing a pilot, overbooking-enabled futures trading (PilotAO) with a DRL-driven adaptive module (AdaptAO). The framework guarantees stability, individual rationality, competitive equilibrium, and weak Pareto optimality through a stable many-to-many matching process and adaptive overbooking updates. Empirical results on real datasets show reduced decision overhead and latency, higher social welfare, and improved resource utilization compared with online spot trading and static futures baselines. The combination of predictive planning and real-time adjustment provides robust performance in dynamic aerial-edge environments, enabling timely and reliable low-altitude mission delivery.
Abstract
Incentive-driven resource trading is essential for UAV applications with intensive, time-sensitive computing demands. Traditional spot trading suffers from negotiation delays and high energy costs, while conventional futures trading struggles to adapt to the dynamic, uncertain UAV-edge environment. To address these challenges, we propose PAST (pilot-and-adaptive stable trading), a novel framework for edge-assisted UAV networks with spatio-temporal dynamism. PAST integrates two complementary mechanisms: PilotAO (pilot trading agreements with overbooking), a risk-aware, overbooking-enabled early-stage decision-making module that establishes long-term, mutually beneficial agreements and boosts resource utilization; and AdaptAO (adaptive trading agreements with overbooking rate update), an intelligent adaptation module that dynamically updates agreements and overbooking rates based on UAV mobility, supply-demand variations, and agreement performance. Together, these mechanisms enable both stability and flexibility, guaranteeing individual rationality, strong stability, competitive equilibrium, and weak Pareto optimality. Extensive experiments on real-world datasets show that PAST consistently outperforms benchmark methods in decision-making overhead, task completion latency, resource utilization, and social welfare. By combining predictive planning with real-time adjustments, PAST offers a valuable reference on robust and adaptive practice for improving low-altitude mission performance.