TimeGNN-Augmented Hybrid-Action MARL for Fine-Grained Task Partitioning and Energy-Aware Offloading in MEC

TL;DR

The paper addresses energy-aware, latency-sensitive task offloading in mobile edge computing (MEC) with intermittently powered edge servers. It introduces TG-DCMADDPG, a framework that fuses TimeGNN-based predictive state augmentation with a discrete–continuous multi-agent DDPG (DC-MADDPG) policy under centralized training and decentralized execution, enabling forward-looking, fine-grained offloading decisions. Empirical results show faster policy convergence, lower energy consumption, reduced latency, and higher task completion rates compared with state-of-the-art baselines, across varying numbers of edge servers and dynamic network conditions. The work demonstrates a scalable, practical approach for proactive MEC management, leveraging spatiotemporal graph modeling to reduce interaction overhead while optimizing energy–latency trade-offs.

Abstract

With the rapid growth of IoT devices and latency-sensitive applications, the demand for both real-time and energy-efficient computing has surged, placing significant pressure on traditional cloud computing architectures. Mobile edge computing (MEC), an emerging paradigm, effectively alleviates the load on cloud centers and improves service quality by offloading computing tasks to edge servers closer to end users. However, the limited computing resources, non-continuous power provisioning (e.g., battery-powered nodes), and highly dynamic systems of edge servers complicate efficient task scheduling and resource allocation. To address these challenges, this paper proposes a multi-agent deep reinforcement learning algorithm, TG-DCMADDPG, and constructs a collaborative computing framework for multiple edge servers, aiming to achieve joint optimization of fine-grained task partitioning and offloading. This approach incorporates a temporal graph neural network (TimeGNN) to model and predict time series of multi-dimensional server state information, thereby reducing the frequency of online interactions and improving policy predictability. Furthermore, a multi-agent deterministic policy gradient algorithm (DC-MADDPG) in a discrete-continuous hybrid action space is introduced to collaboratively optimize task partitioning ratios, transmission power, and priority scheduling strategies. Extensive simulation experiments confirm that TG-DCMADDPG achieves markedly faster policy convergence, superior energy-latency optimization, and higher task completion rates compared with existing state-of-the-art methods, underscoring its robust scalability and practical effectiveness in dynamic and constrained MEC scenarios.

Figures (9)

• Figure 1: Illustrates the considered MEC system, where each edge server operates on battery power without a stable power supply. To prolong service availability and achieve load balancing, a time-series prediction model is employed to forecast the state of edge devices. Based on these predictions, task partitioning and offloading strategies are applied to reduce energy consumption and enhance QoS.
• Figure 2: presents the TG-DCMADDPG framework, which couples TimeGNN-based state prediction with hybrid-action MARL for energy-aware MEC task partitioning. Historical states are transformed into temporal graphs, enabling future server state forecasting. Augmented observations feed a CTDE training loop, where actors output discrete–continuous actions for server selection, task splitting, and power control, while critics guide updates via soft target networks. This integration enables proactive, fine-grained offloading with reduced interaction overhead and improved latency–energy trade-offs in dynamic edge environments.
• Figure 3: Convergence comparison of average cumulative rewards over episodes.
• Figure 4: Impact of the number of edge servers on average cumulative rewards.
• Figure 5: Average energy consumption of different algorithms during training.