A HAPPO based Task Offloading Strategy in Heterogeneous Air-Ground Collaborative MEC Networks

Heterogeneous unmanned aerial vehicle (UAV)assisted air-ground collaborative mobile edge computing (MEC) systems inherently offer enhanced service coverage, reduced latency, and improved resource utilization through coordinated operations of differentiated functional UAVs. However, their heterogeneous resource coordination remains challenged by decoupled optimization of computational resources and airspace management, coupled with strategy conflicts in multi-agent collaboration, leading to suboptimal task completion rates, energy inefficiency, and unbalanced UAV workloads. To address these issues, this paper proposes a Heterogeneous-Agent Proximal Policy Optimization (HAPPO)-based air-ground cooperative MEC task offloading algorithm. By harnessing system heterogeneity as decision-making advantages for agents, a functional mapping mechanism is designed to abstract heterogeneous entities into specialized agents with distinct action and observation spaces. Leveraging HAPPO’s sequential update mechanisms, differential decoupling strategies, and trust region constraints, the algorithm coordinates joint decisions across agents to improve overall system performance. Simulations demonstrate that the proposed algorithm outperforms four baseline algorithms in convergence, task completion rate, UAV load balancing, collision rate, and energy consumption.